nezumi/9router
1
0
Fork 0
mirror of https://github.com/Nezumi-2711/9router.git synced 2026-09-22 13:38:31 +00:00
Code Issues Packages Projects Releases Wiki Activity

merge: sync master into dev

Browse Source 
Sync master into dev while preserving dev-specific feature logic and include the required co-author trailer.

Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
This commit is contained in:
nezumi
2026-07-23 11:52:27 +07:00
co-authored by Copilot
parent 06d6873d52 79918c7830
commit fbbc435bfd
97 changed files with 6384 additions and 985 deletions
Download Patch File Download Diff File Expand all files Collapse all files
CHANGELOG.md
+20 -1
View File
@@ -4,6 +4,25 @@
- **Dokploy**: deploy production Compose from GitHub Actions with serialized rollout tracking, terminal-status handling, health verification, and deployment summaries
- **GitHub Actions**: remove unrelated Docker image publishing and GitBook Pages deployment workflows
# v0.5.40 (2026-07-20)
## Features
- **i18n**: add Khmer (km) translations
- **CLI tools**: configure Grok Build subagent models
- **Kimi**: merge OAuth into dual-auth provider, add K3 / K2.7 models
- **Dashboard**: ProviderTopology flow animation
## Fixes
- **DB**: resolve better-sqlite3 parameter binding crash
- **Translator**: pass `service_tier` through OpenAI → Responses conversion
- **Kiro**: map GPT-5.6 reasoning effort fields
- **Kiro**: validate terminal streams before emitting output
- **Kiro**: map GPT reasoning effort fields
- **Codex**: current `client_version` + refresh-aware model sync
- **Alicode-intl**: split into Coding Plan + Model Studio providers
- **Cursor**: HTTP/2 AgentService support + version bump 3.12.17
- **Dashboard**: cut duplicate API/icon spam, lazy-load provider assets
# v0.5.35 (2026-07-16)
## Fixes
@@ -14,7 +33,7 @@
- **User quota**: show remaining Orbit and Codex headroom in the users table with session and weekly usage details
- **Orbit Provider**: add Anthropic-compatible API-key routing for Claude Opus 4.6–4.8 models
- **xAI**: Grok Imagine video generation (`/v1/videos`) + CLI
- **CLI tools**: Grok Build setup — writes `[model.9router]` to `~/.grok/config.toml`
- **CLI tools**: Grok Build setup — choose separate main/general-purpose/explore/plan models and preserve each model's context window
- **GitHub Copilot**: route Claude models through Copilot's native `/v1/messages`
- **Kiro**: add GPT-5.6 model family (#2596)
- **RTK**: `X-9Router-Token-Saver` header to bypass token savers per request
README.md
+56 -49
View File
@@ -84,7 +84,7 @@ npm install -g 9router
**2. Connect a FREE provider (no signup needed):**
Dashboard → Providers → Connect **Kiro AI** (free Claude unlimited) or **OpenCode Free** (no auth) → Done!
Dashboard → Providers → Connect **Kiro AI** (~50 credits/month free: Claude 4.5 + GLM-5 + MiniMax) or **OpenCode Free** (no auth) → Done!
**3. Use in your CLI tool:**
@@ -127,14 +127,14 @@ Default URLs:
<table>
<tr>
<td align="center" width="320">
<a href="https://www.youtube.com/watch?v=X69n5Lm06Yw">
<img src="https://img.youtube.com/vi/X69n5Lm06Yw/maxresdefault.jpg" alt="Tiết kiệm chi phí LLM với 9Router" width="300"/>
</a><br/>
<b>🇻🇳 Tiếng Việt</b><br/>
<sub>Tiết kiệm chi phí LLM cho OpenClaw với 9Router<br/>by <a href="https://www.youtube.com/c/M%C3%ACAIblog">Mì AI</a></sub>
</td>
<td align="center" width="320">
<td align="center" width="320">
<a href="https://www.youtube.com/watch?v=X69n5Lm06Yw">
<img src="https://img.youtube.com/vi/X69n5Lm06Yw/maxresdefault.jpg" alt="Tiết kiệm chi phí LLM với 9Router" width="300"/>
</a><br/>
<b>🇻🇳 Tiếng Việt</b><br/>
<sub>Tiết kiệm chi phí LLM cho OpenClaw với 9Router<br/>by <a href="https://www.youtube.com/c/M%C3%ACAIblog">Mì AI</a></sub>
</td>
<td align="center" width="320">
<a href="https://youtu.be/VQAw612S27Y">
<img src="https://img.youtube.com/vi/VQAw612S27Y/maxresdefault.jpg" alt="9Router + Claude Code FREE Unlimited Setup" width="300"/>
</a><br/>
@@ -148,10 +148,7 @@ Default URLs:
<b>🇺🇸 English</b><br/>
<sub>9Router + Claude Code FREE Setup<br/>by <a href="https://www.youtube.com/@BuildAIWithHamid">Build AI With Hamid</a></sub>
</td>
</tr>
<tr>
<td align="center" width="320">
<td align="center" width="320">
<a href="https://youtu.be/3dF5GIYMrcQ?si=bAyfyiHbARJQAHj_">
<img src="https://img.youtube.com/vi/3dF5GIYMrcQ/hqdefault.jpg" alt="9Router Setup Tutorial" width="300"/>
</a><br/>
@@ -165,6 +162,8 @@ Default URLs:
<b>🇺🇸 English</b><br/>
<sub>Claude Code FREE Forever — Unlimited Models<br/>by <a href="https://www.youtube.com/@BuildAIWithHamid">Build AI With Hamid</a></sub>
</td>
</tr>
<tr>
<td align="center" width="320">
<a href="https://www.youtube.com/watch?v=Ttpc26m39Dw">
<img src="https://img.youtube.com/vi/Ttpc26m39Dw/maxresdefault.jpg" alt="Claude CLI Free Setup" width="300"/>
@@ -172,10 +171,7 @@ Default URLs:
<b>🇺🇸 English</b><br/>
<sub>Claude CLI Free Setup with 9Router 🚀<br/>by <a href="https://www.youtube.com/@CodeVerseSoban">CodeVerse Soban</a></sub>
</td>
</tr>
<tr>
<td align="center" width="320">
<td align="center" width="320">
<a href="https://www.youtube.com/watch?v=G-5A_D5Pm6Y">
<img src="https://img.youtube.com/vi/G-5A_D5Pm6Y/maxresdefault.jpg" alt="Cài đặt OpenClaw Free A-Z" width="300"/>
</a><br/>
@@ -196,17 +192,15 @@ Default URLs:
<b>🇮🇩 Indonesia</b><br/>
<sub>Koding 24 Jam Anti Rate Limit! Hemat Token AI 65% | Tutorial Quick Setup 9Router 🚀<br/>by <a href="https://www.youtube.com/@krisswuh">Krisswuh</a></sub>
</td>
</tr>
<tr>
<td align="center" width="320">
<td align="center" width="320">
<a href="https://www.youtube.com/watch?v=TXGv4eofe1I">
<img src="https://img.youtube.com/vi/TXGv4eofe1I/mqdefault.jpg" alt="Cara Deploy 9Router di Hugging Face GRATIS Non-Stop! | Alternatif VPS RAM 16GB" width="300"/>
</a><br/>
<b>🇮🇩 Indonesia</b><br/>
<sub>Cara Deploy 9Router di Hugging Face GRATIS Non-Stop! | Alternatif VPS RAM 16GB<br/>by <a href="https://www.youtube.com/@krisswuh">Krisswuh</a></sub>
</td>
</tr>
<tr>
<td align="center" width="320">
<a href="https://www.youtube.com/watch?v=GyX-DLvePW8">
<img src="https://img.youtube.com/vi/GyX-DLvePW8/hqdefault.jpg" alt="این شکلی از هر API ای استفاده کن برای هوش مصنوعی" width="300"/>
@@ -214,8 +208,17 @@ Default URLs:
<b>🇮🇷 Persian-فارسی</b><br/>
<sub dir="rtl">این شکلی از هر API ای استفاده کن برای هوش مصنوعی<br/>by <a href="https://www.youtube.com/@Matin_SenPai">Matin SenPai</a></sub>
</td>
<td align="center" width="320">
<a href="https://www.youtube.com/watch?v=hPusYX-5Pmw">
<img src="https://img.youtube.com/vi/hPusYX-5Pmw/maxresdefault.jpg" alt="Hướng Dẫn Setup OpenClaw + 9Router: Tạo Bot Zalo AI Tự Động Từ A-Z" width="300"/>
</a><br/>
<b>🇻🇳 Tiếng Việt</b><br/>
<sub>Hướng Dẫn Setup OpenClaw + 9Router: Tạo Bot Zalo AI Tự Động Từ A-Z<br/>by <a href="https://github.com/tuanminhhole">tuanminhhole</a></sub>
</td>
<td align="center" width="320"></td>
<td align="center" width="320"></td>
<td align="center" width="320"></td>
</tr>
</table>
</div>
@@ -330,12 +333,12 @@ Default URLs:
<td align="center" width="150">
<img src="./public/providers/kiro.png" width="70" alt="Kiro"/><br/>
<b>Kiro AI</b><br/>
<sub>Claude 4.5 + GLM-5 + MiniMax<br/>Unlimited FREE</sub>
<sub>Claude 4.5 + GLM-5 + MiniMax<br/>50 credits/month free</sub>
</td>
<td align="center" width="150">
<img src="./public/providers/opencode.png" width="70" alt="OpenCode Free"/><br/>
<b>OpenCode Free</b><br/>
<sub>No auth • Auto-fetch models<br/>Unlimited FREE</sub>
<sub>No auth • Auto-fetch models<br/>Free (model list varies)</sub>
</td>
<td align="center" width="150">
<img src="./public/providers/gemini.png" width="70" alt="Vertex AI"/><br/>
@@ -346,7 +349,11 @@ Default URLs:
</table>
</div>
> **Note:** iFlow, Qwen and Gemini CLI free tiers were discontinued in 2026. Use Kiro / OpenCode Free / Vertex instead.
> **Note:** iFlow, Qwen Code and Gemini CLI free tiers were discontinued in 2026. Use Kiro / OpenCode Free / Vertex instead.
>
> **Kiro AI** moved to a paid model in Sep 2025 — the free tier is now capped at **50 credits/month** (plus 500 trial credits for new accounts in the first 30 days). Paid tiers: Pro $20/mo (1,000 credits), Pro+ $40/mo (2,000), Pro Max $100/mo (5,000), Power $200/mo (10,000).
> **OpenCode Free** model list fluctuates over time (some models free only for limited promos) — subject to change without notice.
> **Vertex AI**: the $300 free credit for new GCP accounts is still valid, but since Mar 2026 the **Gemini API endpoint no longer consumes these credits** — call the **Vertex AI Studio** endpoint instead.
### 🔑 API Key Providers (40+)
@@ -600,8 +607,8 @@ Seamless translation between formats:
> The "cost" displayed in Usage Analytics is **for tracking and comparison purposes only**.
> 9Router itself **never charges** you anything. You only pay providers directly (if using paid services).
>
> **Example:** If your dashboard shows "$290 total cost" while using iFlow models, this represents
> what you would have paid using paid APIs directly. Your actual cost = **$0** (iFlow is free unlimited).
> **Example:** If your dashboard shows "$290 total cost" while using Kiro free models, this represents
> what you would have paid using paid APIs directly. Your actual cost = **$0** (Kiro free tier: ~50 credits/mo).
>
> Think of it as a "savings tracker" showing how much you're saving by using free models or
> routing through 9Router!
@@ -629,9 +636,9 @@ Seamless translation between formats:
| **💰 CHEAP** | GLM-5.1 / GLM-4.7 | $0.6/1M | Daily 10AM | Budget backup |
| | MiniMax M2.7 | $0.2/1M | 5-hour rolling | Cheapest option |
| | Kimi K2.5 | $9/mo flat | 10M tokens/mo | Predictable cost |
| **🆓 FREE** | Kiro AI | $0 | Unlimited | Claude 4.5 + GLM-5 + MiniMax free |
| | OpenCode Free | $0 | Unlimited | No auth, auto-fetch models |
| | Vertex AI | $300 credits | New GCP accounts | Gemini 3 Pro + DeepSeek + GLM-5 |
| **🆓 FREE** | Kiro AI | $0 | 50 credits/mo | Claude 4.5 + GLM-5 + MiniMax free (paid tiers above) |
| | OpenCode Free | $0 | Varies* | No auth, auto-fetch models (list changes over time) |
| | Vertex AI | $300 credits | New GCP accounts | Gemini 3 Pro + DeepSeek + GLM-5 (use Vertex AI Studio endpoint for free credits) |
**💡 Pro Tip:** RTK + Kiro AI + OpenCode Free combo = **$0 cost + 20-40% token savings**!
@@ -644,7 +651,7 @@ Seamless translation between formats:
✅ **9Router software = FREE forever** (open source, never charges)
✅ **Dashboard "costs" = Display/tracking only** (not actual bills)
✅ **You pay providers directly** (subscriptions or API fees)
✅ **FREE providers stay FREE** (iFlow, Kiro, Qwen = $0 unlimited)
✅ **FREE providers stay FREE** (Kiro ~50 credits/mo, OpenCode Free, Vertex $300 credits = $0 within free-tier limits) — note iFlow/Qwen/Gemini CLI free tiers were discontinued in 2026
❌ **9Router never sends invoices** or charges your card
**How Cost Display Works:**
@@ -660,7 +667,7 @@ Dashboard Display:
• Display Cost: $290
Reality Check:
• Provider: iFlow (FREE unlimited)
• Provider: Kiro (free tier: ~50 credits/mo)
• Actual Payment: $0.00
• What $290 Means: Amount you SAVED by using free models!
```
@@ -700,7 +707,7 @@ vs. $20 + hitting limits = frustration
```
Combo: "free-forever"
1. kr/claude-sonnet-4.5 (Claude 4.5 free unlimited)
1. kr/claude-sonnet-4.5 (Claude 4.5 free via Kiro, ~50 credits/mo)
2. kr/glm-5 (GLM-5 free via Kiro)
3. oc/<auto> (OpenCode Free, no auth)
@@ -720,7 +727,7 @@ Combo: "always-on"
2. cx/gpt-5.5 (second subscription)
3. glm/glm-5.1 (cheap, resets daily)
4. minimax/MiniMax-M2.7 (cheapest, 5h reset)
5. kr/claude-sonnet-4.5 (free unlimited)
5. kr/claude-sonnet-4.5 (free via Kiro, ~50 credits/mo)
Result: 5 layers of fallback = zero downtime
Monthly cost: $20-200 (subscriptions) + $10-20 (backup)
@@ -754,7 +761,7 @@ The dashboard tracks your token usage and displays **estimated costs** as if you
**Example:**
- **Dashboard shows:** "$290 total cost"
- **Reality:** You're using iFlow (FREE unlimited)
- **Reality:** You're using Kiro free models (~50 credits/mo)
- **Your actual cost:** **$0.00**
- **What $290 means:** Amount you **saved** by using free models instead of paid APIs!
@@ -780,21 +787,21 @@ The cost display is a "savings tracker" to help you understand your usage patter
<details>
<summary><b>🆓 Are FREE providers really unlimited?</b></summary>
**Yes!** The current FREE providers (Kiro, OpenCode Free, Vertex) are genuinely free with **no hidden charges**.
**Mostly!** The current FREE providers (Kiro, OpenCode Free, Vertex) are genuinely free, but free tiers have limits:
These are free services offered by those respective companies:
- **Kiro AI**: Free unlimited Claude 4.5 + GLM-5 + MiniMax via AWS Builder ID / Google / GitHub OAuth
- **OpenCode Free**: No-auth passthrough proxy, models auto-fetched from `opencode.ai/zen/v1/models`
- **Vertex AI**: $300 free credits for new Google Cloud accounts (90 days)
- **Kiro AI**: ~50 credits/month free (plus 500 trial credits for new accounts in the first 30 days) via AWS Builder ID / Google / GitHub OAuth. Paid tiers available above that.
- **OpenCode Free**: No-auth passthrough proxy, models auto-fetched from `opencode.ai/zen/v1/models`. The free model list fluctuates over time (some models free only for limited promos) — subject to change without notice.
- **Vertex AI**: $300 free credits for new Google Cloud accounts (90 days). Since Mar 2026 the Gemini API endpoint no longer consumes these credits — use the **Vertex AI Studio** endpoint instead.
9Router just routes your requests to them - there's no "catch" or future billing. They're truly free services, and 9Router makes them easy to use with fallback support.
9Router just routes your requests to them - there's no "catch" or future billing from 9Router itself. They're truly free services, and 9Router makes them easy to use with fallback support.
**Discontinued free tiers (no longer recommended):**
- ❌ **iFlow**: Was free unlimited, now changed to paid (2026)
- ❌ **Qwen Code**: Free OAuth tier discontinued by Alibaba on 2026-04-15
- ❌ **Gemini CLI**: Still works, but using it with non-CLI tools (Claude, Codex, Cursor...) may result in account bans — only use if you stick to Gemini CLI itself
- ❌ **Qwen Code**: Free OAuth tier fully discontinued by Alibaba on 2026-04-15
- ❌ **Gemini CLI**: Service fully shut down by Google on 2026-06-18 (replaced by the closed-source Antigravity CLI). Discontinued — do not use.
</details>
@@ -806,12 +813,12 @@ These are free services offered by those respective companies:
1. **Start with 100% free combo:**
```
1. gc/gemini-3-flash (180K/month free from Google)
2. if/kimi-k2-thinking (unlimited free from iFlow)
3. qw/qwen3-coder-plus (unlimited free from Qwen)
1. kr/glm-5 (GLM-5 free via Kiro, ~50 credits/mo)
2. OpenCode Free models (no auth, auto-fetched)
3. Vertex AI Gemini 3 Pro (using the Vertex AI Studio endpoint with $300 credits)
```
**Cost: $0/month**
**Cost: $0/month** (within Kiro's free credit cap; OpenCode/Vertex subject to their free-tier limits)
2. **Add cheap backup** only if you need it:
@@ -1036,7 +1043,7 @@ Monthly cost example (100M tokens):
```
Name: free-combo
Models:
1. kr/claude-sonnet-4.5 (Claude 4.5 free unlimited)
1. kr/claude-sonnet-4.5 (Claude 4.5 free via Kiro, ~50 credits/mo)
2. kr/glm-5 (GLM-5 free via Kiro)
3. vertex/gemini-3.1-pro-preview ($300 free credits)
@@ -1301,7 +1308,7 @@ Notes:
- `kimi/kimi-k2.5`
- `kimi/kimi-k2.5-thinking`
**Kiro (`kr/`)** - FREE unlimited:
**Kiro (`kr/`)** - Free (~50 credits/month, paid tiers above):
- `kr/claude-sonnet-4.5`
- `kr/claude-haiku-4.5`
README.zh-CN.md
+30 -26
View File
@@ -82,7 +82,7 @@ npm install -g 9router
**2. 连接免费提供商（无需注册）：**
控制面板 → 提供商 → 连接 **Kiro AI**（免费 Claude 无限量）或 **OpenCode Free**（无需认证）→ 完成！
控制面板 → 提供商 → 连接 **Kiro AI**（约 50 积分/月免费：Claude 4.5 + GLM-5 + MiniMax）或 **OpenCode Free**（无需认证）→ 完成！
**3. 在 CLI 工具中使用：**
@@ -279,12 +279,12 @@ PORT=20128 HOSTNAME=0.0.0.0 NEXT_PUBLIC_BASE_URL=http://localhost:20128 npm run
<td align="center" width="150">
<img src="./public/providers/kiro.png" width="70" alt="Kiro"/><br/>
<b>Kiro AI</b><br/>
<sub>Claude 4.5 + GLM-5 + MiniMax<br/>无限免费</sub>
<sub>Claude 4.5 + GLM-5 + MiniMax<br/>每月 50 积分免费</sub>
</td>
<td align="center" width="150">
<img src="./public/providers/opencode.png" width="70" alt="OpenCode Free"/><br/>
<b>OpenCode Free</b><br/>
<sub>无需认证 • 自动获取模型<br/>无限免费</sub>
<sub>无需认证 • 自动获取模型<br/>免费（模型列表会变）</sub>
</td>
<td align="center" width="150">
<img src="./public/providers/gemini.png" width="70" alt="Vertex AI"/><br/>
@@ -295,7 +295,11 @@ PORT=20128 HOSTNAME=0.0.0.0 NEXT_PUBLIC_BASE_URL=http://localhost:20128 npm run
</table>
</div>
> **注意：** iFlow、Qwen 和 Gemini CLI 的免费等级已于 2026 年停止。请改用 Kiro / OpenCode Free / Vertex。
> **注意：** iFlow、Qwen Code 和 Gemini CLI 的免费等级已于 2026 年停止。请改用 Kiro / OpenCode Free / Vertex。
>
> **Kiro AI** 于 2025 年 9 月转为付费模式 — 免费等级现在上限为**每月 50 积分**（新账户前 30 天另加 500 试用积分）。付费档位：Pro $20/月（1,000 积分）、Pro+ $40/月（2,000）、Pro Max $100/月（5,000）、Power $200/月（10,000）。
> **OpenCode Free** 的模型列表会随时间变化（部分模型仅限时免费）— 可能随时变更，恕不另行通知。
> **Vertex AI**：新 GCP 账户的 $300 免费额度仍然有效，但自 2026 年 3 月起 **Gemini API 端点不再消耗这些额度** — 请改用 **Vertex AI Studio** 端点。
### 🔑 API Key 提供商（40+）
@@ -500,7 +504,7 @@ PORT=20128 HOSTNAME=0.0.0.0 NEXT_PUBLIC_BASE_URL=http://localhost:20128 npm run
> 使用分析中显示的"成本"**仅用于追踪和比较目的**。
> 9Router 本身**永远不会向你收费**。你只直接向提供商付款（如果使用付费服务）。
>
> **示例：** 如果你的控制面板显示使用 iFlow 模型时"总成本 $290"，这代表你如果直接使用付费 API 需要支付的金额。你的实际成本 = **$0**（iFlow 免费无限量）。
> **示例：** 如果你的控制面板显示使用 Kiro 免费模型时"总成本 $290"，这代表你如果直接使用付费 API 需要支付的金额。你的实际成本 = **$0**（Kiro 免费等级：约 50 积分/月）。
>
> 把它想象成一个"节省追踪器"，展示你通过使用免费模型或通过 9Router 路由节省了多少钱！
@@ -527,9 +531,9 @@ PORT=20128 HOSTNAME=0.0.0.0 NEXT_PUBLIC_BASE_URL=http://localhost:20128 npm run
| **💰 低价** | GLM-5.1 / GLM-4.7 | $0.6/1M | 每日 10AM | 预算备份 |
| | MiniMax M2.7 | $0.2/1M | 5小时滚动 | 最便宜选项 |
| | Kimi K2.5 | $9/月固定 | 10M tokens/月 | 可预测成本 |
| **🆓 免费** | Kiro AI | $0 | 无限量 | Claude 4.5 + GLM-5 + MiniMax 免费 |
| | OpenCode Free | $0 | 无限量 | 无需认证，自动获取模型 |
| | Vertex AI | $300 额度 | 新 GCP 账户 | Gemini 3 Pro + DeepSeek + GLM-5 |
| **🆓 免费** | Kiro AI | $0 | 50 积分/月 | Claude 4.5 + GLM-5 + MiniMax 免费（之上为付费档位） |
| | OpenCode Free | $0 | varies* | 无需认证，自动获取模型（列表会变化） |
| | Vertex AI | $300 额度 | 新 GCP 账户 | Gemini 3 Pro + DeepSeek + GLM-5（使用 Vertex AI Studio 端点消耗免费额度） |
**💡 专业提示：** RTK + Kiro AI + OpenCode Free 组合 = **$0 成本 + 节省 20-40% tokens**！
@@ -542,7 +546,7 @@ PORT=20128 HOSTNAME=0.0.0.0 NEXT_PUBLIC_BASE_URL=http://localhost:20128 npm run
✅ **9Router 软件 = 永久免费**（开源，绝不收费）
✅ **控制面板"成本" = 仅用于显示/追踪**（不是实际账单）
✅ **你直接向提供商付款**（订阅或 API 费用）
✅ **免费提供商保持免费**（iFlow、Kiro、Qwen = $0 无限量）
✅ **免费提供商保持免费**（Kiro 约 50 积分/月、OpenCode Free、Vertex $300 额度 = 在免费额度内 $0）— 注意 iFlow/Qwen/Gemini CLI 免费等级已于 2026 年停止
❌ **9Router 永不发送发票** 或扣款
**成本显示如何工作：**
@@ -557,7 +561,7 @@ PORT=20128 HOSTNAME=0.0.0.0 NEXT_PUBLIC_BASE_URL=http://localhost:20128 npm run
• 显示成本：$290
实际检查：
• 提供商：iFlow（免费无限量）
• 提供商：Kiro（免费等级：约 50 积分/月）
• 实际支付：$0.00
• $290 意味着什么：通过使用免费模型节省的金额！
```
@@ -565,7 +569,7 @@ PORT=20128 HOSTNAME=0.0.0.0 NEXT_PUBLIC_BASE_URL=http://localhost:20128 npm run
**付款规则：**
- **订阅提供商**（Claude Code、Codex）：通过他们的网站直接付款
- **低价提供商**（GLM、MiniMax）：直接付款，9Router 只做路由
- **免费提供商**（iFlow、Kiro、Qwen）：真正的永久免费，无隐藏费用
- **免费提供商**（Kiro、OpenCode Free、Vertex）：真正的免费，在免费额度内无隐藏费用
- **9Router**：从不收取任何费用，永远不会
---
@@ -594,7 +598,7 @@ PORT=20128 HOSTNAME=0.0.0.0 NEXT_PUBLIC_BASE_URL=http://localhost:20128 npm run
**解决方案：**
```
组合："free-forever"
1. kr/claude-sonnet-4.5 （Claude 4.5 免费无限量）
1. kr/claude-sonnet-4.5 （通过 Kiro 免费使用 Claude 4.5，约 50 积分/月）
2. kr/glm-5 （通过 Kiro 免费使用 GLM-5）
3. oc/<auto> （OpenCode Free，无需认证）
@@ -613,7 +617,7 @@ PORT=20128 HOSTNAME=0.0.0.0 NEXT_PUBLIC_BASE_URL=http://localhost:20128 npm run
2. cx/gpt-5.5 （第二个订阅）
3. glm/glm-5.1 （低价，每日重置）
4. minimax/MiniMax-M2.7 （最便宜，5小时重置）
5. kr/claude-sonnet-4.5 （免费无限量）
5. kr/claude-sonnet-4.5 （通过 Kiro 免费使用，约 50 积分/月）
结果：5 层切换 = 零停机时间
月成本：$20-200（订阅）+ $10-20（备份）
@@ -645,7 +649,7 @@ PORT=20128 HOSTNAME=0.0.0.0 NEXT_PUBLIC_BASE_URL=http://localhost:20128 npm run
**示例：**
- **控制面板显示：** "$290 总成本"
- **实际情况：** 你在使用 iFlow（免费无限量）
- **实际情况：** 你在使用 Kiro 免费模型（约 50 积分/月）
- **你的实际成本：** **$0.00**
- **$290 的含义：** 你通过使用免费模型而不是付费 API **节省**的金额！
@@ -670,19 +674,19 @@ PORT=20128 HOSTNAME=0.0.0.0 NEXT_PUBLIC_BASE_URL=http://localhost:20128 npm run
<details>
<summary><b>🆓 免费提供商真的是无限量的吗？</b></summary>
**是的！** 当前的免费提供商（Kiro、OpenCode Free、Vertex）是真正的免费，**无隐藏费用**。
**基本上是！** 当前的免费提供商（Kiro、OpenCode Free、Vertex）是真正的免费，但免费等级有上限：
这些是各公司提供的免费服务：
- **Kiro AI**：通过 AWS Builder ID / Google / GitHub OAuth 免费无限量使用 Claude 4.5 + GLM-5 + MiniMax
- **OpenCode Free**：无认证直连代理，模型从 `opencode.ai/zen/v1/models` 自动获取
- **Vertex AI**：新 Google Cloud 账户可获得 $300 免费额度（90 天）
- **Kiro AI**：通过 AWS Builder ID / Google / GitHub OAuth 使用，免费等级约**每月 50 积分**（新账户前 30 天另加 500 试用积分）。之上提供付费档位。
- **OpenCode Free**：无认证直连代理，模型从 `opencode.ai/zen/v1/models` 自动获取。免费模型列表会随时间变化（部分模型仅限时免费）— 可能随时变更。
- **Vertex AI**：新 Google Cloud 账户可获得 $300 免费额度（90 天）。自 2026 年 3 月起 Gemini API 端点不再消耗这些额度 — 请改用 **Vertex AI Studio** 端点。
9Router 只是路由你的请求到它们 — 没有"陷阱"或未来的计费。它们是真正的免费服务，9Router 让它们易于使用并支持切换。
**已停止的免费等级（不再推荐）：**
- ❌ **iFlow**：曾是免费无限量，现在改为付费（2026）
- ❌ **Qwen Code**：阿里巴巴于 2026-04-15 停止免费 OAuth 等级
- ❌ **Gemini CLI**：仍可用，但与非 CLI 工具（Claude、Codex、Cursor...）一起使用可能会导致账户被封 — 仅在你坚持使用 Gemini CLI 本身时才使用
- ❌ **Qwen Code**：阿里巴巴于 2026-04-15 完全停止免费 OAuth 等级
- ❌ **Gemini CLI**：Google 已于 2026-06-18 完全停止服务（由闭源的 Antigravity CLI 取代）。已停止 — 请勿使用。
</details>
@@ -693,11 +697,11 @@ PORT=20128 HOSTNAME=0.0.0.0 NEXT_PUBLIC_BASE_URL=http://localhost:20128 npm run
1. **从 100% 免费组合开始：**
```
1. gc/gemini-3-flash (Google 每月 180K 免费)
2. if/kimi-k2-thinking (iFlow 无限量免费)
3. qw/qwen3-coder-plus (Qwen 无限量免费)
1. kr/glm-5 (通过 Kiro 免费使用 GLM-5，约 50 积分/月)
2. OpenCode Free 模型（无认证，自动获取）
3. Vertex AI Gemini 3 Pro（使用 Vertex AI Studio 端点 + $300 额度）
```
**成本：$0/月**
**成本：$0/月**（在 Kiro 免费积分上限内；OpenCode/Vertex 受各自免费等级限制）
2. **仅在需要时添加低价备份：**
```
@@ -918,7 +922,7 @@ Vertex 合作伙伴（通过 Vertex 提供 Anthropic / DeepSeek / GLM / Qwen）
```
名称：free-combo
模型：
1. kr/claude-sonnet-4.5 (Claude 4.5 免费无限量)
1. kr/claude-sonnet-4.5 (通过 Kiro 免费使用 Claude 4.5，约 50 积分/月)
2. kr/glm-5 (通过 Kiro 免费使用 GLM-5)
3. vertex/gemini-3.1-pro-preview ($300 免费额度)
@@ -1168,7 +1172,7 @@ docker stop 9router && docker rm 9router
- `kimi/kimi-k2.5`
- `kimi/kimi-k2.5-thinking`
**Kiro（`kr/`）** - 免费无限量：
**Kiro（`kr/`）** - 免费（约 50 积分/月，之上为付费档位）：
- `kr/claude-sonnet-4.5`
- `kr/claude-haiku-4.5`
- `kr/glm-5`
cli/cli.js
+196 -12
View File
@@ -3,6 +3,7 @@
const { spawn, exec, execSync } = require("child_process");
const path = require("path");
const fs = require("fs");
const https = require("https");
const net = require("net");
const os = require("os");
@@ -25,6 +26,42 @@ function waitServerReady(port, { timeoutMs = 15000, intervalMs = 150 } = {}) {
});
}
// Native spinner - no external dependency
function createSpinner(text) {
const frames = ["⠋", "⠙", "⠹", "⠸", "⠼", "⠴", "⠦", "⠧", "⠇", "⠏"];
let i = 0;
let interval = null;
let currentText = text;
return {
start() {
if (process.stdout.isTTY) {
process.stdout.write(`\r${frames[0]} ${currentText}`);
interval = setInterval(() => {
process.stdout.write(`\r${frames[i++ % frames.length]} ${currentText}`);
}, 80);
}
return this;
},
stop() {
if (interval) {
clearInterval(interval);
interval = null;
}
if (process.stdout.isTTY) {
process.stdout.write("\r\x1b[K");
}
},
succeed(msg) {
this.stop();
console.log(`✅ ${msg}`);
},
fail(msg) {
this.stop();
console.log(`❌ ${msg}`);
}
};
}
const pkg = require("./package.json");
const { ensureSqliteRuntime, buildEnvWithRuntime } = require("./hooks/sqliteRuntime");
const { ensureTrayRuntime } = require("./hooks/trayRuntime");
@@ -53,6 +90,7 @@ try { ensureTrayRuntime({ silent: true }); } catch {}
// Configuration constants
const APP_NAME = pkg.name; // Use from package.json
const INSTALL_CMD_LATEST = `npm i -g ${APP_NAME}@latest --prefer-online`;
const DEFAULT_PORT = 20128;
const DEFAULT_HOST = "0.0.0.0";
@@ -81,6 +119,7 @@ const PROCESS_IDENTIFIERS = [
let port = DEFAULT_PORT;
let host = DEFAULT_HOST;
let noBrowser = false;
let skipUpdate = false;
let showLog = false;
let trayMode = false;
@@ -95,6 +134,8 @@ for (let i = 0; i < args.length; i++) {
noBrowser = true;
} else if (args[i] === "--log" || args[i] === "-l") {
showLog = true;
} else if (args[i] === "--skip-update") {
skipUpdate = true;
} else if (args[i] === "--tray" || args[i] === "-t") {
trayMode = true;
process.env.TRAY_MODE = "1";
@@ -108,6 +149,7 @@ Options:
-n, --no-browser Don't open browser automatically
-l, --log Show server logs (default: hidden)
-t, --tray Run in system tray mode (background)
--skip-update Skip auto-update check
-h, --help Show this help message
-v, --version Show version
@@ -123,9 +165,26 @@ Commands:
}
}
// Auto-relaunch after update: detached process has no TTY → fallback to tray
if (skipUpdate && !trayMode && !process.stdin.isTTY) {
trayMode = true;
process.env.TRAY_MODE = "1";
}
// Always use Node.js runtime with absolute path
const RUNTIME = process.execPath;
// Compare semver versions: returns 1 if a > b, -1 if a < b, 0 if equal
function compareVersions(a, b) {
const partsA = a.split(".").map(Number);
const partsB = b.split(".").map(Number);
for (let i = 0; i < 3; i++) {
if (partsA[i] > partsB[i]) return 1;
if (partsA[i] < partsB[i]) return -1;
}
return 0;
}
// Get app data dir (matches app/src/lib/dataDir.js convention)
function getAppDataDir() {
return process.platform === "win32"
@@ -150,14 +209,54 @@ function killByPidFile(pidFile) {
} catch { }
}
// Kill tunnel processes (cloudflared/tailscale) by their PID files
function killTunnelByPidFile() {
const tunnelDir = path.join(getAppDataDir(), "tunnel");
killByPidFile(path.join(tunnelDir, "cloudflared.pid"));
killByPidFile(path.join(tunnelDir, "tailscale.pid"));
}
// Kill cloudflared whose --url targets this app's port (covers stale PID file case)
function killCloudflaredByAppPort(appPort) {
if (!appPort) return [];
const portMatchers = [`localhost:${appPort}`, `127.0.0.1:${appPort}`];
const pids = [];
try {
if (process.platform === "win32") {
const psCmd = `powershell -NonInteractive -WindowStyle Hidden -Command "Get-WmiObject Win32_Process -Filter 'Name=\\"cloudflared.exe\\"' | Select-Object ProcessId,CommandLine | ConvertTo-Csv -NoTypeInformation"`;
const output = execSync(psCmd, { encoding: "utf8", windowsHide: true, timeout: 5000 });
const lines = output.split("\n").slice(1).filter(l => l.trim());
lines.forEach(line => {
if (portMatchers.some(m => line.includes(m))) {
const match = line.match(/^"(\d+)"/);
if (match && match[1]) pids.push(match[1]);
}
});
} else {
const output = execSync("ps -eo pid,command 2>/dev/null", { encoding: "utf8", timeout: 5000 });
output.split("\n").forEach(line => {
if (line.includes("cloudflared") && portMatchers.some(m => line.includes(m))) {
const parts = line.trim().split(/\s+/);
const pid = parts[0];
if (pid && !isNaN(pid)) pids.push(pid);
}
});
}
} catch { }
return pids;
}
// Kill all 9router processes
function killAllAppProcesses(appPort) {
return new Promise((resolve) => {
try {
// MITM runs on a separate process and does not block the critical path.
// Background: MITM + tunnel/cloudflared run on separate ports/processes —
// killing them doesn't free the app port, so don't block the critical path.
// Server-side MITM manager has stale-lock recovery and starts deferred (~3s).
setImmediate(() => {
try { killProxyByPidFile(); } catch {}
try { killTunnelByPidFile(); } catch {}
try { killCloudflaredByAppPort(appPort); } catch {}
});
const platform = process.platform;
@@ -356,6 +455,55 @@ function isRestrictedEnvironment() {
return null;
}
// Check if new version available, return latest version or null
function checkForUpdate() {
return new Promise((resolve) => {
if (skipUpdate) {
resolve(null);
return;
}
const spinner = createSpinner("Checking for updates...").start();
let resolved = false;
const safetyTimeout = setTimeout(() => {
if (!resolved) {
resolved = true;
spinner.stop();
resolve(null);
}
}, 8000);
const done = (version) => {
if (resolved) return;
resolved = true;
clearTimeout(safetyTimeout);
spinner.stop();
resolve(version);
};
const req = https.get(`https://registry.npmjs.org/${pkg.name}/latest`, { timeout: 3000 }, (res) => {
let data = "";
res.on("data", chunk => data += chunk);
res.on("end", () => {
try {
const latest = JSON.parse(data);
if (latest.version && compareVersions(latest.version, pkg.version) > 0) {
done(latest.version);
} else {
done(null);
}
} catch (e) {
done(null);
}
});
});
req.on("error", () => done(null));
req.on("timeout", () => { req.destroy(); done(null); });
});
}
// Open browser
function openBrowser(url) {
const platform = process.platform;
@@ -390,24 +538,39 @@ if (!fs.existsSync(serverPath)) {
process.exit(1);
}
// Start server immediately; run update check in parallel (not on the critical path).
const updatePromise = checkForUpdate();
killAllAppProcesses(port)
.then(() => killProcessOnPort(port))
.then(() => startServer());
.then(() => startServer(updatePromise));
// Show interface selection menu
async function showInterfaceMenu() {
async function showInterfaceMenu(latestVersion) {
const { selectMenu } = require("./src/cli/utils/input");
const { clearScreen } = require("./src/cli/utils/display");
const { getEndpoint } = require("./src/cli/utils/endpoint");
clearScreen();
const displayHost = getDisplayHost();
const serverUrl = `http://${displayHost}:${port}`;
// Detect tunnel/local mode for server URL display
let serverUrl;
try {
const { endpoint, tunnelEnabled } = await getEndpoint(port);
serverUrl = tunnelEnabled ? endpoint.replace(/\/v1$/, "") : `http://${displayHost}:${port}`;
} catch (e) {
serverUrl = `http://${displayHost}:${port}`;
}
const subtitle = `🚀 Server: \x1b[32m${serverUrl}\x1b[0m`;
const menuItems = [];
if (latestVersion) {
menuItems.push({ label: `Update to v${latestVersion} (current: v${pkg.version})`, icon: "⬆" });
}
menuItems.push(
{ label: "Web UI (Open in Browser)", icon: "🌐" },
{ label: "Terminal UI (Interactive CLI)", icon: "💻" },
@@ -417,16 +580,21 @@ async function showInterfaceMenu() {
const selected = await selectMenu(`Choose Interface (v${pkg.version})`, menuItems, 0, subtitle);
if (selected === 0) return "web";
if (selected === 1) return "terminal";
if (selected === 2) return "hide";
const offset = latestVersion ? 1 : 0;
if (latestVersion && selected === 0) return "update";
if (selected === offset) return "web";
if (selected === offset + 1) return "terminal";
if (selected === offset + 2) return "hide";
return "exit";
}
const MAX_RESTARTS = 2;
const RESTART_RESET_MS = 30000; // Reset counter if alive > 30s
function startServer() {
function startServer(updatePromise) {
// Accept either a Promise (parallel update check) or a resolved value.
const latestVersionPromise = Promise.resolve(updatePromise);
const displayHost = getDisplayHost();
const url = `http://${displayHost}:${port}/dashboard`;
// Surface real network exposure when bound to all interfaces (default 0.0.0.0).
@@ -444,7 +612,7 @@ function startServer() {
function spawnServer() {
serverStartTime = Date.now();
crashLog = [];
const child = spawn(RUNTIME, ["--max-old-space-size=6144", serverPath], {
const child = spawn(RUNTIME, ["--dns-result-order=ipv4first", "--max-old-space-size=6144", serverPath], {
cwd: standaloneDir,
stdio: showLog ? "inherit" : ["ignore", "ignore", "pipe"],
detached: true,
@@ -480,6 +648,8 @@ function startServer() {
} catch (e) { }
// Kill MIT server (privileged process) via PID file
killProxyByPidFile();
// Kill cloudflared/tailscale via PID file (only this app's tunnel)
killTunnelByPidFile();
// Kill server process directly
if (server.pid) {
process.kill(server.pid, "SIGKILL");
@@ -556,14 +726,28 @@ function startServer() {
// Wait for server to be ready, then show interface menu loop + tray
waitServerReady(port).then(async () => {
// Resolve parallel update check (already running); don't block server start on it.
const latestVersion = await latestVersionPromise;
// Start tray icon alongside TUI
initTrayIcon();
try {
while (true) {
const choice = await showInterfaceMenu();
const choice = await showInterfaceMenu(latestVersion);
if (choice === "web") {
if (choice === "update") {
isShuttingDown = true;
const { clearScreen } = require("./src/cli/utils/display");
clearScreen();
console.log(`\n⬆ Update v${pkg.version} → v${latestVersion}\n`);
console.log(`Run this after exit:\n`);
console.log(` \x1b[33m${INSTALL_CMD_LATEST}\x1b[0m\n`);
cleanup();
await killAllAppProcesses(port);
await killProcessOnPort(port);
setTimeout(() => process.exit(0), 200);
return;
} else if (choice === "web") {
openBrowser(url);
// Wait for user to come back
const { pause } = require("./src/cli/utils/input");
@@ -601,7 +785,7 @@ function startServer() {
// Windows/Linux: spawn detached bgProcess (systray works fine in child)
console.log(`\n⏳ Starting background process... (tray icon will appear in ~3s)`);
const bgProcess = spawn(process.execPath, [__filename, "--tray", "-p", port.toString()], {
const bgProcess = spawn(process.execPath, ["--dns-result-order=ipv4first", __filename, "--tray", "--skip-update", "-p", port.toString()], {
detached: true,
stdio: "ignore",
windowsHide: true,
cli/package.json
+1 -1
View File
@@ -1,6 +1,6 @@
{
"name": "9router",
"version": "0.5.35",
"version": "0.5.40",
"description": "9Router CLI - Start and manage 9Router server",
"bin": {
"9router": "./cli.js"
cli/src/cli/menus/providers.js
+1 -1
View File
@@ -131,7 +131,7 @@ const APIKEY_PROVIDERS = {
openrouter: { id: "openrouter", name: "OpenRouter" },
glm: { id: "glm", name: "GLM Coding" },
minimax: { id: "minimax", name: "Minimax Coding" },
kimi: { id: "kimi", name: "Kimi Coding" },
kimi: { id: "kimi", name: "Kimi" },
openai: { id: "openai", name: "OpenAI" },
anthropic: { id: "anthropic", name: "Anthropic" },
gemini: { id: "gemini", name: "Gemini" },
open-sse/config/appConstants.js
+39 -6
View File
@@ -1,6 +1,7 @@
import { platform, arch } from "os";
import { platform, arch, hostname } from "os";
import { PROVIDERS, PROVIDER_OAUTH } from "./providers.js";
import { ANTIGRAVITY_IDE_USER_AGENT } from "../providers/shared.js";
import { createRequire } from "module";
// === Gemini CLI === derive từ registry gemini-cli.transport
export const GEMINI_CLI_VERSION = PROVIDERS["gemini-cli"]?.cliVersion;
@@ -171,12 +172,44 @@ export const OAUTH_ENDPOINTS = {
github: { token: PROVIDER_OAUTH["github"]?.tokenUrl, auth: PROVIDER_OAUTH["github"]?.authorizeUrl, deviceCode: PROVIDER_OAUTH["github"]?.deviceCodeUrl },
};
// Generate Kimi OAuth custom headers
export function buildKimiHeaders() {
let _appVersion;
function getAppPackageVersion() {
if (_appVersion) return _appVersion;
try {
const require = createRequire(import.meta.url);
_appVersion = require("../../package.json").version || "0.0.0";
} catch {
_appVersion = process.env.npm_package_version || "0.0.0";
}
return _appVersion;
}
// Kimi Code OAuth / API headers (CLIProxyAPI internal/auth/kimi commonHeaders parity).
// deviceId must stay stable per connection for the whole OAuth session.
export function buildKimiHeaders(deviceId) {
const osName = platform();
const architecture = arch();
let deviceModel = `${osName} ${architecture}`;
if (osName === "darwin") deviceModel = `macOS ${architecture}`;
else if (osName === "win32") deviceModel = `Windows ${architecture}`;
else if (osName === "linux") deviceModel = `Linux ${architecture}`;
let deviceName = "unknown";
try {
deviceName = hostname() || "unknown";
} catch {
deviceName = "unknown";
}
const resolvedId = (typeof deviceId === "string" && deviceId.trim())
? deviceId.trim()
: `kimi-${Date.now()}`;
return {
"X-Msh-Platform": "9router",
"X-Msh-Version": "2.1.2",
"X-Msh-Device-Model": typeof process !== "undefined" ? `${process.platform} ${process.arch}` : "unknown",
"X-Msh-Device-Id": `kimi-${Date.now()}`
"X-Msh-Version": getAppPackageVersion(),
"X-Msh-Device-Name": deviceName,
"X-Msh-Device-Model": deviceModel,
"X-Msh-Device-Id": resolvedId,
};
}
open-sse/config/kiroConstants.js
+48 -11
View File
@@ -8,8 +8,8 @@
* - `-agentic` model suffix detection + chunked-write system prompt
* - reasoning / thinking trigger detection (Anthropic-Beta header,
* Claude `thinking`, OpenAI `reasoning_effort`, AMP/Cursor magic tag)
* - the `<thinking_mode>enabled</thinking_mode>` system-prompt injection
* that turns Kiro reasoning on
* - schema-specific native effort fields for supported GPT and Claude models
* - legacy `<thinking_mode>` system-prompt injection for other models
*
* Kiro upstream does not advertise `-agentic` model IDs; they are a 9router
* fiction. The suffix is stripped before the request leaves this process.
@@ -109,6 +109,7 @@ export function resolveKiroThinkingBudget(body, headers, model) {
const cfg = extractThinking(body);
if (cfg) {
if (cfg.mode === "none") return null;
if (cfg.mode === "level" && cfg.level === "disabled") return null;
if (cfg.mode === "budget") return cfg.budget;
if (cfg.mode === "level") return effortToBudget(cfg.level) ?? KIRO_THINKING_BUDGET_DEFAULT;
return KIRO_THINKING_BUDGET_DEFAULT;
@@ -144,9 +145,30 @@ export function extractKiroEffortLevel(body) {
return null;
}
export function buildKiroAdditionalModelRequestFields(body) {
const effort = extractKiroEffortLevel(body);
function extractKiroGptEffortLevel(body) {
const effort =
body?.output_config?.effort ??
body?.reasoning_effort ??
(typeof body?.reasoning === "object" ? body.reasoning?.effort : null);
if (typeof effort !== "string") return null;
const normalized = effort.toLowerCase();
if (normalized === "max") return "xhigh";
// Kiro CLI does not advertise an explicit GPT "none" wire value; omit it.
if (["low", "medium", "high", "xhigh"].includes(normalized)) {
return normalized;
}
return null;
}
export function buildKiroAdditionalModelRequestFields(body, effortPath = "output_config") {
const effort = effortPath === "reasoning"
? extractKiroGptEffortLevel(body)
: extractKiroEffortLevel(body);
if (!effort) return undefined;
if (effortPath === "reasoning") {
// Mirrors Kiro CLI/KAS buildEffortRequestFields("reasoning") for GPT.
return { reasoning: { effort } };
}
// Mirrors Kiro CLI/KAS buildEffortRequestFields("output_config").
return {
thinking: { type: "adaptive", display: "summarized" },
@@ -154,12 +176,15 @@ export function buildKiroAdditionalModelRequestFields(body) {
};
}
export function supportsKiroAdditionalModelRequestFields(model) {
if (typeof model !== "string") return false;
export function resolveKiroEffortPath(model) {
if (typeof model !== "string") return null;
const normalized = model.toLowerCase().replace(/-/g, ".");
if (!normalized.includes("claude")) return false;
if (/(?:^|[/.])gpt[/.]5[/.]6(?:[/.]|$)/.test(normalized)) {
return "reasoning";
}
if (!normalized.includes("claude")) return null;
const match = normalized.match(/(?:^|[/.])claude(?:[/.][a-z]+)*[/.](\d+)(?:[/.](\d+))?(?:[/.]|$)/);
if (!match) return false;
if (!match) return null;
const [, majorText, minorText] = match;
const major = Number(majorText);
const minor = minorText === undefined ? null : Number(minorText);
@@ -167,12 +192,24 @@ export function supportsKiroAdditionalModelRequestFields(model) {
// Kiro rejected additionalModelRequestFields on legacy 4.5 models in live smoke.
// Default future Claude/Kiro models to supported so new model releases do not
// need a code allowlist update.
return !(major < 4 || (major === 4 && (minor === null || minor <= 5 || dateSuffixMinor)));
return major < 4 || (major === 4 && (minor === null || minor <= 5 || dateSuffixMinor))
? null
: "output_config";
}
export function supportsKiroAdditionalModelRequestFields(model) {
return resolveKiroEffortPath(model) !== null;
}
export function usesKiroNativeGptEffort(body, model) {
return resolveKiroEffortPath(model) === "reasoning"
&& extractKiroGptEffortLevel(body) !== null;
}
export function buildKiroAdditionalModelRequestFieldsForModel(body, model) {
if (!supportsKiroAdditionalModelRequestFields(model)) return undefined;
return buildKiroAdditionalModelRequestFields(body);
const effortPath = resolveKiroEffortPath(model);
if (!effortPath) return undefined;
return buildKiroAdditionalModelRequestFields(body, effortPath);
}
/**
open-sse/executors/cursor.js
+415 -1
View File
@@ -1,8 +1,11 @@
import { BaseExecutor } from "./base.js";
import { PROVIDERS } from "../config/providers.js";
import { PROVIDERS, PROVIDER_OAUTH } from "../config/providers.js";
import { HTTP_STATUS } from "../config/runtimeConfig.js";
import {
generateCursorBody,
encodeField,
wrapConnectRPCFrame,
decodeMessage,
parseConnectRPCFrame,
extractTextFromResponse
} from "../utils/cursorProtobuf.js";
@@ -13,6 +16,7 @@ import { chatChunkSse } from "../utils/sse.js";
import { FORMATS } from "../translator/formats.js";
import { proxyAwareFetch } from "../utils/proxyFetch.js";
import zlib from "zlib";
import crypto from "crypto";
// Detect cloud environment
const isCloudEnv = () => {
@@ -38,6 +42,130 @@ const COMPRESS_FLAG = {
GZIP_TRAILER: 0x03
};
const AGENT_RUN_PATH = "/agent.v1.AgentService/Run";
const PROTOBUF_LEN = 2;
const PROTOBUF_VARINT = 0;
function concatBuffers(...parts) {
const length = parts.reduce((total, part) => total + part.length, 0);
const result = new Uint8Array(length);
let offset = 0;
for (const part of parts) {
result.set(part, offset);
offset += part.length;
}
return result;
}
const agentString = (field, value) => encodeField(field, PROTOBUF_LEN, value);
const agentMessage = (field, value) => encodeField(field, PROTOBUF_LEN, value);
const agentBool = (field, value) => encodeField(field, PROTOBUF_VARINT, value ? 1 : 0);
function textFromContent(content) {
if (typeof content === "string") return content;
if (!Array.isArray(content)) return "";
return content
.filter((part) => part?.type === "text" && typeof part.text === "string")
.map((part) => part.text)
.join("\n");
}
function isAgentTextRequest(body) {
// Many compatible clients always attach their built-in tool schemas, even
// for a normal text turn. Cursor's retired ChatService rejects those
// requests; AgentService can still answer the text turn, so ignore schemas
// here. A real tool-call/result conversation is kept on the legacy path
// until its AgentService tool protocol is implemented.
return Array.isArray(body?.messages) && body.messages.every((message) => {
if (message?.tool_calls?.length || message?.role === "tool") return false;
return typeof message?.content === "string"
|| Array.isArray(message?.content) && message.content.every((part) => part?.type === "text");
});
}
function encodeHistoryMessage(message) {
const content = textFromContent(message?.content);
if (!content) return null;
// ConversationHistoryMessage.user / .assistant -> repeated content -> text.
const text = agentString(1, content);
if (message.role === "assistant") {
return agentMessage(2, agentMessage(1, agentMessage(1, text)));
}
return agentMessage(1, agentMessage(1, agentMessage(1, text)));
}
function buildAgentRunFrame(messages, model) {
const system = messages
.filter((message) => message?.role === "system")
.map((message) => textFromContent(message.content))
.filter(Boolean)
.join("\n\n");
const chatMessages = messages.filter((message) => message?.role !== "system");
const currentIndex = [...chatMessages].map((message) => message?.role).lastIndexOf("user");
const current = currentIndex >= 0 ? chatMessages[currentIndex] : chatMessages.at(-1);
const history = chatMessages
.slice(0, currentIndex >= 0 ? currentIndex : -1)
.map(encodeHistoryMessage)
.filter(Boolean);
const userText = textFromContent(current?.content) || "Continue.";
// agent.v1.UserMessageAction.user_message and its optional history.
const userMessage = concatBuffers(
agentString(1, userText),
agentString(2, crypto.randomUUID()),
);
const conversationHistory = history.length
? concatBuffers(...history.map((entry) => agentMessage(1, entry)))
: null;
const userAction = concatBuffers(
agentMessage(1, userMessage),
...(conversationHistory ? [agentMessage(7, conversationHistory)] : []),
);
const conversationAction = agentMessage(1, userAction);
const requestedModel = concatBuffers(agentString(1, model), agentBool(7, true));
const runRequest = concatBuffers(
// An empty ConversationStateStructure starts a fresh local agent session.
agentMessage(1, new Uint8Array()),
agentMessage(2, conversationAction),
...(system ? [agentString(8, system)] : []),
agentMessage(9, requestedModel),
);
// agent.v1.AgentClientMessage.run_request.
return wrapConnectRPCFrame(agentMessage(1, runRequest));
}
function extractAgentString(message, field) {
const value = message?.get(field)?.[0]?.value;
return value ? Buffer.from(value).toString("utf8") : "";
}
function decodeAgentFrames(buffer, onFrame) {
let pending = Buffer.from(buffer || []);
while (pending.length >= 5) {
const flags = pending[0];
const length = pending.readUInt32BE(1);
if (pending.length < 5 + length) break;
let payload = pending.subarray(5, 5 + length);
pending = pending.subarray(5 + length);
if (flags & COMPRESS_FLAG.GZIP) {
payload = zlib.gunzipSync(payload);
}
if (!(flags & COMPRESS_FLAG.TRAILER)) onFrame(payload);
}
return pending;
}
function createRequestContextResponse() {
// AgentService asks every run for client context. 9router has no IDE file
// context, so acknowledge with an empty RequestContext.
const requestContextSuccess = agentMessage(1, new Uint8Array());
const requestContextResult = agentMessage(1, requestContextSuccess);
const execClientMessage = agentMessage(10, requestContextResult);
return wrapConnectRPCFrame(agentMessage(2, execClientMessage));
}
const CURSOR_STREAM_DEBUG = process.env.CURSOR_STREAM_DEBUG === "1";
const debugLog = (...args) => {
if (CURSOR_STREAM_DEBUG) console.log(...args);
@@ -253,7 +381,293 @@ export class CursorExecutor extends BaseExecutor {
});
}
/**
* AgentService (agent.api5.cursor.sh) is HTTP/2-only. Node's fetch/undici speaks
* HTTP/1.1 and fails with HTTPParserError on the h2 preface — use http2 duplex.
*/
openAgentHttp2Stream(url, headers, signal) {
if (!http2) {
throw new Error("HTTP/2 is required for Cursor AgentService (endpoint is h2-only)");
}
const urlObj = new URL(url);
const client = http2.connect(`https://${urlObj.host}`);
const chunkQueue = [];
let waiting = null;
let ended = false;
let streamError = null;
let req = null;
const wake = (result) => {
if (!waiting) return;
const resolve = waiting;
waiting = null;
resolve(result);
};
const fail = (error) => {
if (streamError) return;
streamError = error;
ended = true;
wake(null);
};
const close = () => {
try { req?.destroy(); } catch {}
try { client.close(); } catch {}
};
client.on("error", fail);
req = client.request({
":method": "POST",
":path": urlObj.pathname,
":authority": urlObj.host,
":scheme": "https",
...headers,
});
req.on("error", fail);
req.on("data", (chunk) => {
if (waiting) wake({ value: chunk, done: false });
else chunkQueue.push(chunk);
});
req.on("end", () => {
ended = true;
wake({ value: undefined, done: true });
});
if (signal) {
const onAbort = () => {
fail(new Error("Request aborted"));
close();
};
if (signal.aborted) onAbort();
else signal.addEventListener("abort", onAbort, { once: true });
}
const responseHeaders = new Promise((resolve, reject) => {
const onEarlyError = (error) => reject(error);
client.once("error", onEarlyError);
req.once("error", onEarlyError);
req.once("response", (hdrs) => {
client.off("error", onEarlyError);
req.off("error", onEarlyError);
resolve(hdrs);
});
});
return {
responseHeaders,
write(frame) {
if (req && !req.destroyed) req.write(Buffer.from(frame));
},
end() {
try { if (req && !req.destroyed) req.end(); } catch {}
},
close,
async read() {
if (chunkQueue.length) return { value: chunkQueue.shift(), done: false };
if (ended) {
if (streamError) throw streamError;
return { value: undefined, done: true };
}
const result = await new Promise((resolve) => { waiting = resolve; });
if (streamError) throw streamError;
return result || { value: undefined, done: true };
},
};
}
async executeAgent({ model, body, stream, credentials, signal }) {
const agentEndpoint = PROVIDER_OAUTH.cursor?.agentEndpoint;
if (!agentEndpoint) throw new Error("Cursor AgentService endpoint is not configured");
const url = `${agentEndpoint}${AGENT_RUN_PATH}`;
const headers = this.buildHeaders(credentials);
const requestController = new AbortController();
if (signal?.addEventListener) {
signal.addEventListener("abort", () => requestController.abort(signal.reason), { once: true });
}
let session;
try {
session = this.openAgentHttp2Stream(url, headers, requestController.signal);
session.write(buildAgentRunFrame(body.messages || [], model));
} catch (error) {
throw new Error(`Cursor AgentService request failed: ${error.message}`);
}
let responseHeaders;
try {
responseHeaders = await session.responseHeaders;
} catch (error) {
session.close();
throw new Error(`Cursor AgentService request failed: ${error.message}`);
}
const status = Number(responseHeaders[":status"] || 0);
if (status !== 200) {
let errorText = "";
try {
while (true) {
const { done, value } = await session.read();
if (done) break;
errorText += Buffer.from(value).toString("utf8");
}
} catch {}
session.close();
return {
response: new Response(JSON.stringify({
error: { message: `Cursor AgentService ${status}: ${errorText || "request failed"}`, type: "api_error" },
}), { status: status || HTTP_STATUS.SERVER_ERROR, headers: { "Content-Type": "application/json" } }),
url,
headers,
transformedBody: body,
responseFormat: FORMATS.OPENAI,
};
}
// The Claude SSE translator derives Anthropic's message ID by stripping
// `chatcmpl-`. Keep the remaining ID in Anthropic's required `msg_` form
// so strict clients such as Claude Code accept the completed stream.
const responseId = `chatcmpl-msg_${Date.now()}`;
const created = Math.floor(Date.now() / 1000);
let pending = Buffer.alloc(0);
let finished = false;
const consume = async (onEvent) => {
try {
while (!finished) {
const { done, value } = await session.read();
if (done) break;
pending = Buffer.concat([pending, Buffer.from(value)]);
pending = decodeAgentFrames(pending, (payload) => {
const serverMessage = decodeMessage(payload);
// agent.v1.AgentServerMessage.interaction_update
if (serverMessage.has(1)) {
const update = decodeMessage(serverMessage.get(1)[0].value);
if (update.has(1)) {
const textDelta = extractAgentString(decodeMessage(update.get(1)[0].value), 1);
if (textDelta) onEvent({ type: "text", value: textDelta });
}
// Cursor's AgentService emits internal reasoning without the
// cryptographic signature required by Anthropic thinking blocks.
// Forwarding it makes strict Anthropic clients (Claude Code)
// discard or wait on an otherwise complete response. Keep the
// reasoning upstream-only and emit the normal answer text.
if (update.has(14)) {
finished = true;
onEvent({ type: "done" });
}
}
// AgentService requests IDE context before producing a response.
// Return an empty context; 9router is not coupled to an editor.
if (serverMessage.has(2)) {
const execRequest = decodeMessage(serverMessage.get(2)[0].value);
if (execRequest.has(10)) {
session.write(createRequestContextResponse());
} else {
finished = true;
onEvent({ type: "error", value: "Cursor AgentService requested an unsupported IDE tool" });
onEvent({ type: "done" });
}
}
});
}
} finally {
try { session.end(); } catch {}
try { session.close(); } catch {}
if (!finished) onEvent({ type: "done" });
}
};
if (stream === false) {
let content = "";
let reasoning = "";
let agentError = null;
await consume((event) => {
if (event.type === "text") content += event.value;
else if (event.type === "thinking") reasoning += event.value;
else if (event.type === "error") agentError = event.value;
});
if (agentError) {
return {
response: new Response(JSON.stringify({ error: { message: agentError, type: "api_error" } }), {
status: HTTP_STATUS.BAD_REQUEST,
headers: { "Content-Type": "application/json" },
}),
url,
headers,
transformedBody: body,
responseFormat: FORMATS.OPENAI,
};
}
return {
response: new Response(JSON.stringify({
id: responseId,
object: "chat.completion",
created,
model,
choices: [{ index: 0, message: { role: "assistant", content: content || null, ...(reasoning ? { reasoning_content: reasoning } : {}) }, finish_reason: "stop" }],
usage: estimateUsage(body, content.length, FORMATS.OPENAI),
}), { headers: { "Content-Type": "application/json" } }),
url,
headers,
transformedBody: body,
responseFormat: FORMATS.OPENAI,
};
}
const encoder = new TextEncoder();
const responseStream = new ReadableStream({
start(controller) {
consume((event) => {
if (event.type === "text") {
controller.enqueue(encoder.encode(chatChunkSse({ id: responseId, created, model, delta: { content: event.value } })));
} else if (event.type === "thinking") {
controller.enqueue(encoder.encode(chatChunkSse({ id: responseId, created, model, delta: { reasoning_content: event.value } })));
} else if (event.type === "error") {
controller.enqueue(encoder.encode(chatChunkSse({ id: responseId, created, model, delta: { content: `\n[${event.value}]` } })));
} else if (event.type === "done") {
controller.enqueue(encoder.encode(chatChunkSse({ id: responseId, created, model, delta: {}, finishReason: "stop" })));
controller.enqueue(encoder.encode(SSE_DONE));
controller.close();
}
}).catch((error) => controller.error(error));
},
cancel() {
requestController.abort();
},
});
return {
response: new Response(responseStream, { headers: SSE_HEADERS }),
url,
headers,
transformedBody: body,
responseFormat: FORMATS.OPENAI,
};
}
async execute({ model, body, stream, credentials, signal, log, proxyOptions = null }) {
if (isAgentTextRequest(body)) {
try {
return await this.executeAgent({ model, body, stream, credentials, signal });
} catch (error) {
return {
response: new Response(JSON.stringify({
error: { message: error.message, type: "connection_error", code: "" },
}), { status: HTTP_STATUS.SERVER_ERROR, headers: { "Content-Type": "application/json" } }),
url: `${PROVIDER_OAUTH.cursor?.agentEndpoint || ""}${AGENT_RUN_PATH}`,
headers: {},
transformedBody: body,
};
}
}
const url = this.buildUrl();
const headers = this.buildHeaders(credentials);
const transformedBody = this.transformRequest(model, body, stream, credentials);
open-sse/executors/default.js
+12 -6
View File
@@ -38,7 +38,8 @@ function applyAuth(headers, desc, credentials) {
// Provider-specific header quirks kept as small hooks (not pure auth).
const HEADER_HOOKS = {
kimiHeaders: (h) => Object.assign(h, buildKimiHeaders()),
// Stable device_id from OAuth connection (CLIProxyAPI KimiTokenStorage.DeviceID)
kimiHeaders: (h, c) => Object.assign(h, buildKimiHeaders(c?.providerSpecificData?.deviceId)),
clineHeaders: (h, c) => Object.assign(h, buildClineHeaders(c.apiKey || c.accessToken)),
kilocodeOrg: (h, c) => { if (c.providerSpecificData?.orgId) h["X-Kilocode-OrganizationID"] = c.providerSpecificData.orgId; },
claudeOverlay: (h) => {
@@ -227,7 +228,8 @@ export class DefaultExecutor extends BaseExecutor {
kiro: () => this.refreshKiro(credentials.refreshToken, proxyOptions),
cline: () => this.refreshCline(credentials.refreshToken, proxyOptions),
clinepass: () => this.refreshCline(credentials.refreshToken, proxyOptions),
"kimi-coding": () => this.refreshKimiCoding(credentials.refreshToken, proxyOptions),
kimi: () => this.refreshKimi(credentials, proxyOptions),
"kimi-coding": () => this.refreshKimi(credentials, proxyOptions),
kilocode: () => this.refreshKilocode(credentials.refreshToken, proxyOptions)
};
@@ -307,16 +309,20 @@ export class DefaultExecutor extends BaseExecutor {
return { accessToken, refreshToken: data?.refreshToken || refreshToken, expiresIn };
}
async refreshKimiCoding(refreshToken, proxyOptions = null) {
const kimiHeaders = buildKimiHeaders();
const response = await proxyAwareFetch(PROVIDERS["kimi-coding"].refreshUrl, {
// CLIProxyAPI DeviceFlowClient.RefreshToken — form body + X-Msh-* headers + stable device_id
async refreshKimi(credentials, proxyOptions = null) {
const refreshToken = credentials.refreshToken;
const cfg = PROVIDERS.kimi || PROVIDERS["kimi-coding"];
if (!cfg?.refreshUrl || !cfg?.clientId) return null;
const kimiHeaders = buildKimiHeaders(credentials?.providerSpecificData?.deviceId);
const response = await proxyAwareFetch(cfg.refreshUrl, {
method: "POST",
headers: {
"Content-Type": "application/x-www-form-urlencoded",
"Accept": "application/json",
...kimiHeaders
},
body: new URLSearchParams({ grant_type: "refresh_token", refresh_token: refreshToken, client_id: PROVIDERS["kimi-coding"].clientId })
body: new URLSearchParams({ grant_type: "refresh_token", refresh_token: refreshToken, client_id: cfg.clientId })
}, proxyOptions);
if (!response.ok) return null;
const tokens = await response.json();
open-sse/executors/kiro.js
+1031 -422
View File
File diff suppressed because it is too large Load Diff
open-sse/handlers/chatCore.js
+11 -3
View File
@@ -291,12 +291,16 @@ export async function handleChatCore({ body, modelInfo, credentials, log, onCred
// Execute request
let providerResponse, providerUrl, providerHeaders, finalBody;
// Most executors return their registry format. Cursor AgentService is an
// exception: it is decoded by the executor into OpenAI-compatible output.
let providerResponseFormat = targetFormat;
try {
const result = await executor.execute({ model, body: translatedBody, stream, credentials, signal: streamController.signal, log, proxyOptions });
providerResponse = result.response;
providerUrl = result.url;
providerHeaders = result.headers;
finalBody = result.transformedBody;
providerResponseFormat = result.responseFormat || targetFormat;
reqLogger.logTargetRequest(providerUrl, providerHeaders, finalBody);
} catch (error) {
trackPendingRequest(model, provider, connectionId, false, true);
@@ -335,7 +339,11 @@ export async function handleChatCore({ body, modelInfo, credentials, log, onCred
}
try {
const retryResult = await executor.execute({ model, body: translatedBody, stream, credentials, signal: streamController.signal, log, proxyOptions });
if (retryResult.response.ok) { providerResponse = retryResult.response; providerUrl = retryResult.url; }
if (retryResult.response.ok) {
providerResponse = retryResult.response;
providerUrl = retryResult.url;
providerResponseFormat = retryResult.responseFormat || targetFormat;
}
} catch { log?.warn?.("TOKEN", `${provider.toUpperCase()} | retry after refresh failed`); }
} else {
log?.warn?.("TOKEN", `${provider.toUpperCase()} | refresh failed`);
@@ -382,14 +390,14 @@ export async function handleChatCore({ body, modelInfo, credentials, log, onCred
// True non-streaming response
if (!stream) {
const result = await handleNonStreamingResponse({ ...sharedCtx, providerResponse, sourceFormat, targetFormat, reqLogger, toolNameMap, trackDone, appendLog });
const result = await handleNonStreamingResponse({ ...sharedCtx, providerResponse, sourceFormat, targetFormat: providerResponseFormat, reqLogger, toolNameMap, trackDone, appendLog });
streamController.handleComplete();
return result;
}
// Streaming response
const { onStreamComplete, streamDetailId } = buildOnStreamComplete({ ...sharedCtx });
return handleStreamingResponse({ ...sharedCtx, providerResponse, sourceFormat, targetFormat, userAgent, reqLogger, toolNameMap, streamController, onStreamComplete, streamDetailId });
return handleStreamingResponse({ ...sharedCtx, providerResponse, sourceFormat, targetFormat: providerResponseFormat, userAgent, reqLogger, toolNameMap, streamController, onStreamComplete, streamDetailId });
}
export function isTokenExpiringSoon(expiresAt, bufferMs = 5 * 60 * 1000) {
open-sse/handlers/chatCore/sseToJsonHandler.js
+13 -1
View File
@@ -40,15 +40,21 @@ function pickAssistantMessageForChatCompletion(output) {
*/
export function parseSSEToOpenAIResponse(rawSSE, fallbackModel) {
const chunks = [];
let streamError = null;
for (const line of String(rawSSE || "").split("\n")) {
const trimmed = line.trim();
if (!trimmed.startsWith("data:")) continue;
const payload = trimmed.slice(5).trim();
if (!payload || payload === "[DONE]") continue;
try { chunks.push(JSON.parse(payload)); } catch { /* ignore malformed lines */ }
try {
const chunk = JSON.parse(payload);
if (chunk?.error) streamError = chunk.error;
else chunks.push(chunk);
} catch { /* ignore malformed lines */ }
}
if (streamError) return { error: streamError };
if (chunks.length === 0) return null;
const first = chunks[0];
@@ -196,6 +202,12 @@ export async function handleForcedSSEToJson({ providerResponse, sourceFormat, pr
const sseText = await providerResponse.text();
const parsed = parseSSEToOpenAIResponse(sseText, model);
if (!parsed) return createErrorResult(HTTP_STATUS.BAD_GATEWAY, "Invalid SSE response for non-streaming request");
if (parsed.error) {
return createErrorResult(
HTTP_STATUS.BAD_GATEWAY,
parsed.error.message || "Upstream SSE stream failed"
);
}
if (onRequestSuccess) await onRequestSuccess();
open-sse/providers/capabilities.js
+24 -1
View File
@@ -96,10 +96,23 @@ export const MODEL_CAPABILITIES = {
// Qwen plain coder/text (no vision) — registry "vision-model" / "coder-model" aliases
"vision-model": { vision: true, reasoning: true, thinkingFormat: "qwen", contextWindow: 1000000 },
"coder-model": { reasoning: true, thinkingFormat: "qwen", contextWindow: 1000000 },
// Kimi flagship + coding (platform + Kimi Code ids) — vision/video native
"kimi-k3": { vision: true, videoInput: true, reasoning: true, thinkingFormat: "kimi", thinkingCanDisable: false, contextWindow: 1048576, maxOutput: 131072 },
"k3": { vision: true, videoInput: true, reasoning: true, thinkingFormat: "kimi", thinkingCanDisable: false, contextWindow: 1048576, maxOutput: 131072 },
"kimi-for-coding": { vision: true, videoInput: true, reasoning: true, thinkingFormat: "kimi", thinkingCanDisable: false, contextWindow: 262144, maxOutput: 65536 },
"kimi-for-coding-highspeed": { vision: true, videoInput: true, reasoning: true, thinkingFormat: "kimi", thinkingCanDisable: false, contextWindow: 262144, maxOutput: 65536 },
"kimi-k2.7-code": { vision: true, videoInput: true, reasoning: true, thinkingFormat: "kimi", thinkingCanDisable: false, contextWindow: 262144, maxOutput: 65536 },
"kimi-k2.7-code-highspeed": { vision: true, videoInput: true, reasoning: true, thinkingFormat: "kimi", thinkingCanDisable: false, contextWindow: 262144, maxOutput: 65536 },
};
const KIRO_GPT_5_6_CAPABILITIES = { vision: true, reasoning: true, search: true, thinkingFormat: "openai", contextWindow: 272000, maxOutput: 128000 };
// Codex OAuth (ChatGPT backend) — per-model context window reported by upstream
// (lower than OpenAI API's 1.05M). Sol differs from Terra/Luna. #2720
const CODEX_GPT_56_SOL_CAPS = { vision: true, reasoning: true, search: true, thinkingFormat: "openai", contextWindow: 372000, maxOutput: 128000 };
const CODEX_GPT_56_DEFAULT_CAPS = { vision: true, reasoning: true, search: true, thinkingFormat: "openai", contextWindow: 272000, maxOutput: 128000 };
/**
* Provider-specific capability overrides. Keyed by provider alias/id.
*/
@@ -113,6 +126,14 @@ export const PROVIDER_CAPABILITIES = {
"deepseek-ai/deepseek-v4-pro": { reasoning: true, thinkingFormat: "openai", contextWindow: 1000000, maxOutput: 65536 },
"deepseek-ai/deepseek-v4-flash": { reasoning: true, thinkingFormat: "openai", contextWindow: 1000000, maxOutput: 65536 },
},
"codex": {
"gpt-5.6-sol": CODEX_GPT_56_SOL_CAPS,
"gpt-5.6-sol-review": CODEX_GPT_56_SOL_CAPS,
"gpt-5.6-terra": CODEX_GPT_56_DEFAULT_CAPS,
"gpt-5.6-terra-review": CODEX_GPT_56_DEFAULT_CAPS,
"gpt-5.6-luna": CODEX_GPT_56_DEFAULT_CAPS,
"gpt-5.6-luna-review": CODEX_GPT_56_DEFAULT_CAPS,
},
"kiro": {
"gpt-5.6-sol": KIRO_GPT_5_6_CAPABILITIES,
"gpt-5.6-terra": KIRO_GPT_5_6_CAPABILITIES,
@@ -222,7 +243,9 @@ export const PATTERN_CAPABILITIES = [
{ pattern: "*qwen*", caps: { reasoning: true, thinkingFormat: "qwen", contextWindow: 262144 } },
// ── Kimi (enabled→reasoning_effort; K2.7-code cannot disable) ─────
{ pattern: "*kimi*k2.7*code*", caps: { vision: true, reasoning: true, thinkingFormat: "kimi", thinkingCanDisable: false, contextWindow: 262144, maxOutput: 262144 } },
{ pattern: "*kimi*k3*", caps: { vision: true, videoInput: true, reasoning: true, thinkingFormat: "kimi", thinkingCanDisable: false, contextWindow: 1048576, maxOutput: 131072 } },
{ pattern: "*kimi*for-coding*", caps: { vision: true, videoInput: true, reasoning: true, thinkingFormat: "kimi", thinkingCanDisable: false, contextWindow: 262144, maxOutput: 65536 } },
{ pattern: "*kimi*k2.7*code*", caps: { vision: true, videoInput: true, reasoning: true, thinkingFormat: "kimi", thinkingCanDisable: false, contextWindow: 262144, maxOutput: 65536 } },
{ pattern: "*kimi*k2*", caps: { vision: true, reasoning: true, thinkingFormat: "kimi", contextWindow: 262144, maxOutput: 262144 } },
{ pattern: "*kimi*", caps: { reasoning: true, thinkingFormat: "kimi", contextWindow: 262144 } },
open-sse/providers/pricing.js
+9
View File
@@ -75,10 +75,18 @@ export const MODEL_PRICING = {
"qwen3-coder-flash": { input: 0.50, output: 2.00, cached: 0.25, reasoning: 3.00, cache_creation: 0.50 },
// === Kimi ===
// Official platform.kimi.ai: cache-hit / cache-miss / output per 1M tokens
"kimi-k3": { input: 3.00, output: 15.00, cached: 0.30, reasoning: 15.00, cache_creation: 3.00 },
"k3": { input: 3.00, output: 15.00, cached: 0.30, reasoning: 15.00, cache_creation: 3.00 },
"kimi-k2.7-code": { input: 0.95, output: 4.00, cached: 0.19, reasoning: 4.00, cache_creation: 0.95 },
"kimi-k2.7-code-highspeed": { input: 1.90, output: 8.00, cached: 0.38, reasoning: 8.00, cache_creation: 1.90 },
"kimi-for-coding": { input: 0.95, output: 4.00, cached: 0.19, reasoning: 4.00, cache_creation: 0.95 },
"kimi-for-coding-highspeed": { input: 1.90, output: 8.00, cached: 0.38, reasoning: 8.00, cache_creation: 1.90 },
"kimi-k2": { input: 1.00, output: 4.00, cached: 0.50, reasoning: 6.00, cache_creation: 1.00 },
"kimi-k2-thinking": { input: 1.50, output: 6.00, cached: 0.75, reasoning: 9.00, cache_creation: 1.50 },
"kimi-k2.5": { input: 1.20, output: 4.80, cached: 0.60, reasoning: 7.20, cache_creation: 1.20 },
"kimi-k2.5-thinking": { input: 1.80, output: 7.20, cached: 0.90, reasoning: 10.80, cache_creation: 1.80 },
"kimi-k2.6": { input: 1.00, output: 4.00, cached: 0.50, reasoning: 6.00, cache_creation: 1.00 },
"kimi-latest": { input: 1.00, output: 4.00, cached: 0.50, reasoning: 6.00, cache_creation: 1.00 },
// === DeepSeek ===
@@ -185,6 +193,7 @@ export const PATTERN_PRICING = [
// --- Kimi ---
{ pattern: "kimi-*-thinking", pricing: { input: 1.80, output: 7.20, cached: 0.90, reasoning: 10.80, cache_creation: 1.80 } },
{ pattern: "kimi-k3*", pricing: { input: 3.00, output: 15.00, cached: 0.30, reasoning: 15.00, cache_creation: 3.00 } },
{ pattern: "kimi-k2*", pricing: { input: 1.20, output: 4.80, cached: 0.60, reasoning: 7.20, cache_creation: 1.20 } },
{ pattern: "kimi-*", pricing: { input: 1.00, output: 4.00, cached: 0.50, reasoning: 6.00, cache_creation: 1.00 } },
open-sse/providers/registry/alicode-intl.js
+4 -4
View File
@@ -3,18 +3,18 @@ export default {
priority: 10,
alias: "alicode-intl",
display: {
name: "Alibaba Intl",
name: "Alibaba Coding",
icon: "cloud",
color: "#FF6A00",
textIcon: "ALi",
website: "https://modelstudio.console.alibabacloud.com",
website: "https://www.alibabacloud.com/product/coding",
notice: {
apiKeyUrl: "https://modelstudio.console.alibabacloud.com/?apiKey=1",
apiKeyUrl: "https://www.alibabacloud.com/product/coding",
},
},
category: "apikey",
transport: {
baseUrl: "https://dashscope-intl.aliyuncs.com/compatible-mode/v1/chat/completions",
baseUrl: "https://coding-intl.dashscope.aliyuncs.com/v1/chat/completions",
headers: {},
quirks: { preserveCacheControl: true },
},
open-sse/providers/registry/alims-intl.js
+32
View File
@@ -0,0 +1,32 @@
// Model Studio Intl — standard DashScope API keys (sk-...), NOT Coding Plan keys.
// Sibling of alicode-intl (Coding Plan). Two key types use two different hosts.
export default {
id: "alims-intl",
priority: 11,
alias: "alims-intl",
display: {
name: "Alibaba Studio",
icon: "cloud",
color: "#FF6A00",
textIcon: "ALi",
website: "https://modelstudio.console.alibabacloud.com",
notice: {
apiKeyUrl: "https://modelstudio.console.alibabacloud.com/?apiKey=1",
},
},
category: "apikey",
transport: {
baseUrl: "https://dashscope-intl.aliyuncs.com/compatible-mode/v1/chat/completions",
headers: {},
quirks: { preserveCacheControl: true },
},
models: [
{ id: "qwen3.5-plus", name: "Qwen3.5 Plus" },
{ id: "kimi-k2.5", name: "Kimi K2.5" },
{ id: "glm-5", name: "GLM 5" },
{ id: "MiniMax-M2.5", name: "MiniMax M2.5" },
{ id: "qwen3-coder-next", name: "Qwen3 Coder Next" },
{ id: "qwen3-coder-plus", name: "Qwen3 Coder Plus" },
{ id: "glm-4.7", name: "GLM 4.7" },
],
};
open-sse/providers/registry/cursor.js
+3 -3
View File
@@ -23,7 +23,7 @@ export default {
"Content-Type": "application/connect+proto",
"User-Agent": "connect-es/1.6.1",
},
clientVersion: "3.1.0",
clientVersion: "3.12.17",
},
models: [
{ id: "default", name: "Auto (Server Picks)" },
@@ -44,11 +44,11 @@ export default {
oauth: {
apiEndpoint: "https://api2.cursor.sh",
chatEndpoint: "/aiserver.v1.ChatService/StreamUnifiedChatWithTools",
modelsEndpoint: "/aiserver.v1.AiService/GetDefaultModelNudgeData",
modelsEndpoint: "/agent.v1.AgentService/GetUsableModels",
api3Endpoint: "https://api3.cursor.sh",
agentEndpoint: "https://agent.api5.cursor.sh",
agentNonPrivacyEndpoint: "https://agentn.api5.cursor.sh",
clientVersion: "3.1.0",
clientVersion: "3.12.17",
clientType: "ide",
dbKeys: {
accessToken: "cursorAuth/accessToken",
open-sse/providers/registry/index.js
+48 -48
View File
@@ -52,53 +52,53 @@ import p49 from "./jina-ai.js";
import p50 from "./jina-reader.js";
import p51 from "./kilocode.js";
import p52 from "./kimchi.js";
import p53 from "./kimi-coding.js";
import p54 from "./kimi.js";
import p55 from "./kiro.js";
import p56 from "./linkup.js";
import p57 from "./local-device.js";
import p58 from "./mimo-free.js";
import p59 from "./minimax-cn.js";
import p60 from "./minimax.js";
import p61 from "./mistral.js";
import p62 from "./mmf.js";
import p63 from "./nanobanana.js";
import p64 from "./nebius.js";
import p65 from "./nvidia.js";
import p66 from "./ollama-local.js";
import p67 from "./ollama.js";
import p68 from "./openai.js";
import p69 from "./opencode-go.js";
import p70 from "./opencode.js";
import p71 from "./openrouter.js";
import p72 from "./perplexity-web.js";
import p73 from "./perplexity.js";
import p74 from "./perplexity-agent.js";
import p75 from "./playht.js";
import p76 from "./qoder.js";
import p77 from "./qwen.js";
import p78 from "./recraft.js";
import p79 from "./runwayml.js";
import p80 from "./sdwebui.js";
import p81 from "./searchapi.js";
import p82 from "./searxng.js";
import p83 from "./serper.js";
import p84 from "./siliconflow.js";
import p85 from "./stability-ai.js";
import p86 from "./tavily.js";
import p87 from "./together.js";
import p88 from "./topaz.js";
import p89 from "./tortoise.js";
import p90 from "./venice.js";
import p91 from "./vercel-ai-gateway.js";
import p92 from "./vertex-partner.js";
import p93 from "./vertex.js";
import p94 from "./volcengine-ark.js";
import p95 from "./voyage-ai.js";
import p96 from "./xai.js";
import p97 from "./xiaomi-mimo.js";
import p98 from "./xiaomi-tokenplan.js";
import p99 from "./youcom.js";
import p53 from "./kimi.js";
import p54 from "./kiro.js";
import p55 from "./linkup.js";
import p56 from "./local-device.js";
import p57 from "./mimo-free.js";
import p58 from "./minimax-cn.js";
import p59 from "./minimax.js";
import p60 from "./mistral.js";
import p61 from "./mmf.js";
import p62 from "./nanobanana.js";
import p63 from "./nebius.js";
import p64 from "./nvidia.js";
import p65 from "./ollama-local.js";
import p66 from "./ollama.js";
import p67 from "./openai.js";
import p68 from "./opencode-go.js";
import p69 from "./opencode.js";
import p70 from "./openrouter.js";
import p71 from "./perplexity-web.js";
import p72 from "./perplexity.js";
import p73 from "./perplexity-agent.js";
import p74 from "./playht.js";
import p75 from "./qoder.js";
import p76 from "./qwen.js";
import p77 from "./recraft.js";
import p78 from "./runwayml.js";
import p79 from "./sdwebui.js";
import p80 from "./searchapi.js";
import p81 from "./searxng.js";
import p82 from "./serper.js";
import p83 from "./siliconflow.js";
import p84 from "./stability-ai.js";
import p85 from "./tavily.js";
import p86 from "./together.js";
import p87 from "./topaz.js";
import p88 from "./tortoise.js";
import p89 from "./venice.js";
import p90 from "./vercel-ai-gateway.js";
import p91 from "./vertex-partner.js";
import p92 from "./vertex.js";
import p93 from "./volcengine-ark.js";
import p94 from "./voyage-ai.js";
import p95 from "./xai.js";
import p96 from "./xiaomi-mimo.js";
import p97 from "./xiaomi-tokenplan.js";
import p98 from "./youcom.js";
import p99 from "./alims-intl.js";
import p100 from "./orbit-provider.js";
export default [
@@ -202,5 +202,5 @@ export default [
p97,
p98,
p99,
p100
p100,
];
open-sse/providers/registry/kimi-coding.js
-65
View File
@@ -1,65 +0,0 @@
import { CLAUDE_API_HEADERS, KIMI_CODING_BASE_URL } from "../shared.js";
export default {
id: "kimi-coding",
hidden: true,
priority: 120,
alias: "kmc",
display: {
name: "Kimi Coding",
icon: "psychology",
color: "#1E40AF",
textIcon: "KC",
website: "https://kimi.moonshot.cn",
notice: {
signupUrl: "https://kimi.moonshot.cn",
},
},
category: "oauth",
transport: {
baseUrl: "https://api.kimi.com/coding/v1/messages",
format: "claude",
urlSuffix: "?beta=true",
headers: { ...CLAUDE_API_HEADERS },
clientId: "17e5f671-d194-4dfb-9706-5516cb48c098",
tokenUrl: "https://auth.kimi.com/api/oauth/token",
refreshUrl: "https://auth.kimi.com/api/oauth/token",
auth: {
combined: true,
header: "x-api-key",
scheme: "raw",
hooks: [
"kimiHeaders",
],
},
},
// Multi-endpoint: pick the transport matching client sourceFormat to skip translation.
transports: [
{
format: "openai",
baseUrl: "https://api.kimi.com/coding/v1/chat/completions",
auth: { combined: true, header: "Authorization", scheme: "bearer", hooks: ["kimiHeaders"] },
},
{
format: "claude",
baseUrl: "https://api.kimi.com/coding/v1/messages",
urlSuffix: "?beta=true",
headers: { ...CLAUDE_API_HEADERS },
auth: { combined: true, header: "x-api-key", scheme: "raw", hooks: ["kimiHeaders"] },
},
],
models: [
{ id: "kimi-k2.6", name: "Kimi K2.6" },
{ id: "kimi-k2.5", name: "Kimi K2.5" },
{ id: "kimi-k2.5-thinking", name: "Kimi K2.5 Thinking" },
{ id: "kimi-latest", name: "Kimi Latest" },
],
oauth: {
deviceCodeUrl: "https://auth.kimi.com/api/oauth/device_authorization",
tokenUrl: "https://auth.kimi.com/api/oauth/token",
refreshLeadMs: 300000,
},
features: {
usage: true,
},
};
open-sse/providers/registry/kimi.js
+40 -7
View File
@@ -1,9 +1,14 @@
import { CLAUDE_API_HEADERS, KIMI_CODING_BASE_URL } from "../shared.js";
import { CLAUDE_API_HEADERS } from "../shared.js";
// Dual auth (same pattern as xai): OAuth = Kimi Code subscription (device code),
// API key = platform.moonshot / api.kimi.com. Transport is shared.
// CLIProxyAPI parity: client_id, auth.kimi.com device+token, X-Msh-* headers, device_id.
export default {
id: "kimi",
priority: 170,
alias: "kimi",
// Legacy id + short alias from former kimi-coding registry entry
aliases: ["kimi-coding", "kmc"],
display: {
name: "Kimi",
icon: "psychology",
@@ -12,18 +17,25 @@ export default {
website: "https://kimi.moonshot.cn",
notice: {
apiKeyUrl: "https://platform.moonshot.ai/console/api-keys",
signupUrl: "https://www.kimi.com/code",
},
},
category: "apikey",
category: "oauth",
authModes: ["oauth", "apikey"],
hasOAuth: true,
transport: {
baseUrl: "https://api.kimi.com/coding/v1/messages",
format: "claude",
urlSuffix: "?beta=true",
headers: { ...CLAUDE_API_HEADERS },
clientId: "17e5f671-d194-4dfb-9706-5516cb48c098",
tokenUrl: "https://auth.kimi.com/api/oauth/token",
refreshUrl: "https://auth.kimi.com/api/oauth/token",
auth: {
combined: true,
header: "x-api-key",
scheme: "raw",
hooks: ["kimiHeaders"],
},
},
// Multi-endpoint: pick the transport matching client sourceFormat to skip translation.
@@ -31,26 +43,47 @@ export default {
{
format: "openai",
baseUrl: "https://api.kimi.com/coding/v1/chat/completions",
auth: { combined: true, header: "Authorization", scheme: "bearer" },
auth: { combined: true, header: "Authorization", scheme: "bearer", hooks: ["kimiHeaders"] },
},
{
format: "claude",
baseUrl: "https://api.kimi.com/coding/v1/messages",
urlSuffix: "?beta=true",
headers: { ...CLAUDE_API_HEADERS },
auth: { combined: true, header: "x-api-key", scheme: "raw" },
auth: { combined: true, header: "x-api-key", scheme: "raw", hooks: ["kimiHeaders"] },
},
],
models: [
// Flagship K3 — platform.kimi.ai id `kimi-k3`, Kimi Code OAuth id `k3` (up to 1M)
{ id: "kimi-k3", name: "Kimi K3" },
{ id: "k3", name: "Kimi K3 (Code)" },
// Kimi Code subscription stable ids (map to K2.7 Code backend)
{ id: "kimi-for-coding", name: "Kimi for Coding" },
{ id: "kimi-for-coding-highspeed", name: "Kimi for Coding Highspeed" },
// Pay-as-you-go platform ids
{ id: "kimi-k2.7-code", name: "Kimi K2.7 Code" },
{ id: "kimi-k2.7-code-highspeed", name: "Kimi K2.7 Code Highspeed" },
{ id: "kimi-k2.6", name: "Kimi K2.6" },
{ id: "kimi-k2.5", name: "Kimi K2.5" },
{ id: "kimi-k2.5-thinking", name: "Kimi K2.5 Thinking" },
{ id: "kimi-latest", name: "Kimi Latest" },
],
serviceKinds: ["llm","webSearch"],
serviceKinds: ["llm", "webSearch"],
searchViaChat: {
defaultModel: "kimi-k2.5",
defaultModel: "kimi-k3",
endpoint: "https://api.moonshot.cn/v1/chat/completions",
pricingUrl: "https://platform.moonshot.ai/docs/pricing/chat",
pricingUrl: "https://platform.kimi.ai/docs/pricing/chat",
},
oauth: {
clientId: "17e5f671-d194-4dfb-9706-5516cb48c098",
deviceCodeUrl: "https://auth.kimi.com/api/oauth/device_authorization",
tokenUrl: "https://auth.kimi.com/api/oauth/token",
refreshUrl: "https://auth.kimi.com/api/oauth/token",
// CLIProxyAPI refreshThresholdSeconds = 300
refreshLeadMs: 300000,
authorizeDeviceUrl: "https://www.kimi.com/code/authorize_device",
},
features: {
usage: true,
},
};
open-sse/services/cursorModels.js
+187
View File
@@ -0,0 +1,187 @@
/**
* Cursor live model catalog fetcher.
*
* Cursor exposes the account-specific model picker through the AgentService
* `GetUsableModels` Connect RPC. Unlike the static provider registry, this
* includes models newly enabled for the account and omits unavailable ones.
*/
import crypto from "crypto";
import http2 from "http2";
import { PROVIDER_OAUTH } from "../providers/index.js";
import { buildCursorHeaders } from "../utils/cursorChecksum.js";
import { decodeMessage } from "../utils/cursorProtobuf.js";
const FETCH_TIMEOUT_MS = 10_000;
const CACHE_TTL_MS = 5 * 60 * 1000;
// agent.v1.ModelDetails protobuf field numbers.
const MODEL_ID_FIELD = 1;
const DISPLAY_MODEL_ID_FIELD = 3;
const DISPLAY_NAME_FIELD = 4;
const DISPLAY_NAME_SHORT_FIELD = 5;
const RESPONSE_MODELS_FIELD = 1;
/** @type {Map<string, { expiresAt: number, models: { id: string, name: string }[] }>} */
const catalogCache = new Map();
function getCursorModelsUrl() {
const config = PROVIDER_OAUTH.cursor;
if (!config?.agentEndpoint || !config?.modelsEndpoint) return null;
return `${config.agentEndpoint.replace(/\/$/, "")}${config.modelsEndpoint}`;
}
function cacheKey(credentials) {
const seed = [
credentials?.providerSpecificData?.machineId,
credentials?.accessToken,
].filter(Boolean).join(":");
if (!seed) return "cursor-anonymous";
return crypto.createHash("sha256").update(`cursor:${seed}`).digest("hex");
}
function firstString(fields, fieldNumber) {
const value = fields.get(fieldNumber)?.[0]?.value;
if (!value || typeof value === "number") return "";
return Buffer.from(value).toString("utf8");
}
/**
* Decode Cursor's `agent.v1.GetUsableModelsResponse` protobuf payload.
* The response contains repeated `agent.v1.ModelDetails` messages in field 1.
*/
export function parseCursorUsableModels(payload) {
const response = decodeMessage(payload);
const seen = new Set();
const models = [];
for (const entry of response.get(RESPONSE_MODELS_FIELD) || []) {
if (!entry?.value || typeof entry.value === "number") continue;
const detail = decodeMessage(entry.value);
const id = firstString(detail, MODEL_ID_FIELD).trim();
if (!id || seen.has(id)) continue;
seen.add(id);
const name = (
firstString(detail, DISPLAY_NAME_FIELD)
|| firstString(detail, DISPLAY_NAME_SHORT_FIELD)
|| firstString(detail, DISPLAY_MODEL_ID_FIELD)
|| id
).trim();
models.push({ id, name });
}
return models;
}
/**
* agent.api5.cursor.sh is HTTP/2-only; Node fetch/undici cannot speak h2.
* Unary GetUsableModels uses an unframed protobuf body (application/proto).
*/
function http2PostProto(url, headers, body, signal, timeoutMs) {
return new Promise((resolve, reject) => {
const urlObj = new URL(url);
const client = http2.connect(`https://${urlObj.host}`);
const chunks = [];
let responseHeaders = {};
let settled = false;
const finish = (fn) => (...args) => {
if (settled) return;
settled = true;
clearTimeout(timeoutId);
try { client.close(); } catch {}
fn(...args);
};
const timeoutId = setTimeout(finish(() => {
reject(new Error("Cursor GetUsableModels timed out"));
}), timeoutMs);
client.on("error", finish(reject));
const req = client.request({
":method": "POST",
":path": urlObj.pathname,
":authority": urlObj.host,
":scheme": "https",
...headers,
});
req.on("response", (hdrs) => { responseHeaders = hdrs; });
req.on("data", (chunk) => { chunks.push(chunk); });
req.on("end", finish(() => {
resolve({
status: Number(responseHeaders[":status"] || 0),
body: Buffer.concat(chunks),
});
}));
req.on("error", finish(reject));
if (signal) {
const onAbort = finish(() => reject(new Error("Request aborted")));
if (signal.aborted) onAbort();
else signal.addEventListener("abort", onAbort, { once: true });
}
req.end(body && body.length ? Buffer.from(body) : undefined);
});
}
async function fetchCursorCatalog(credentials, signal) {
const accessToken = credentials?.accessToken;
const machineId = credentials?.providerSpecificData?.machineId;
const url = getCursorModelsUrl();
if (!accessToken || !machineId || !url) return null;
const headers = {
...buildCursorHeaders(accessToken, machineId, credentials?.providerSpecificData?.ghostMode !== false),
// Connect unary calls use an unframed protobuf body, unlike Cursor chat's
// streaming `application/connect+proto` endpoint.
accept: "application/proto",
"content-type": "application/proto",
};
delete headers["connect-accept-encoding"];
delete headers["connect-protocol-version"];
const response = await http2PostProto(url, headers, new Uint8Array(), signal, FETCH_TIMEOUT_MS);
if (response.status !== 200) {
const error = new Error(`Cursor GetUsableModels returned ${response.status}`);
error.status = response.status;
throw error;
}
return parseCursorUsableModels(new Uint8Array(response.body));
}
/**
* Resolve the live Cursor catalog for the authenticated account.
* Returns null on any failure so callers can fall back to static models.
*/
export async function resolveCursorModels(credentials, options = {}) {
if (!credentials?.accessToken || !credentials?.providerSpecificData?.machineId) {
options.log?.debug?.("CURSOR_MODELS", "No Cursor access token or machine ID; skipping live fetch");
return null;
}
const key = cacheKey(credentials);
const now = Date.now();
if (!options.forceRefresh) {
const cached = catalogCache.get(key);
if (cached?.expiresAt > now) return { models: cached.models };
}
try {
const models = await fetchCursorCatalog(credentials, options.signal);
if (!models?.length) return null;
catalogCache.set(key, { expiresAt: now + CACHE_TTL_MS, models });
return { models };
} catch (error) {
options.log?.warn?.("CURSOR_MODELS", `Live model fetch failed: ${error?.message || error}`);
return null;
}
}
export function clearCursorModelCache() {
catalogCache.clear();
}
open-sse/services/tokenRefresh.js
+9 -1
View File
@@ -3,6 +3,7 @@ import { OAUTH_ENDPOINTS, REFRESH_LEAD_MS } from "../config/appConstants.js";
import {
refreshXaiToken,
refreshAccessToken,
refreshKimiToken,
refreshClaudeOAuthToken,
refreshGoogleToken,
refreshQwenToken,
@@ -18,6 +19,7 @@ import {
// Re-export all provider refresh functions (preserves public API for all consumers)
export {
refreshAccessToken,
refreshKimiToken,
refreshClaudeOAuthToken,
refreshGoogleToken,
refreshQwenToken,
@@ -44,7 +46,10 @@ export function isUnrecoverableRefreshError(result) {
}
export function getRefreshLeadMs(provider) {
return REFRESH_LEAD_MS[provider] || TOKEN_EXPIRY_BUFFER_MS;
if (REFRESH_LEAD_MS[provider]) return REFRESH_LEAD_MS[provider];
// Legacy id after kimi-coding → kimi merge
if (provider === "kimi-coding" && REFRESH_LEAD_MS.kimi) return REFRESH_LEAD_MS.kimi;
return TOKEN_EXPIRY_BUFFER_MS;
}
export function parseVertexSaJson(apiKey) {
@@ -133,6 +138,9 @@ const REFRESH_HANDLERS = {
"grok-cli": (c, log) => refreshXaiToken(c.refreshToken, log),
gcli: (c, log) => refreshXaiToken(c.refreshToken, log),
"codebuddy-cn": (c, log) => refreshCodebuddyToken(c.refreshToken, log),
// Kimi Code OAuth (merged into id `kimi`); legacy id still routes here
kimi: (c, log) => refreshKimiToken(c.refreshToken, c, log),
"kimi-coding": (c, log) => refreshKimiToken(c.refreshToken, c, log),
vertex: vertexRefreshHandler,
"vertex-partner": vertexRefreshHandler
};
open-sse/services/tokenRefresh/providers.js
+47 -1
View File
@@ -1,5 +1,5 @@
import { PROVIDERS, PROVIDER_OAUTH } from "../../config/providers.js";
import { OAUTH_ENDPOINTS, GITHUB_COPILOT } from "../../config/appConstants.js";
import { OAUTH_ENDPOINTS, GITHUB_COPILOT, buildKimiHeaders } from "../../config/appConstants.js";
import { proxyAwareFetch } from "../../utils/proxyFetch.js";
import { dedupRefresh } from "./dedup.js";
import { buildExternalIdpRefreshParams } from "../../../src/lib/oauth/kiroExternalIdp.js";
@@ -91,6 +91,52 @@ export async function refreshAccessToken(provider, refreshToken, credentials, lo
}, log);
}
// CLIProxyAPI DeviceFlowClient.RefreshToken: form body (no client_secret) + X-Msh-* headers
export async function refreshKimiToken(refreshToken, credentials, log) {
const config = PROVIDERS.kimi;
if (!config?.refreshUrl || !config?.clientId) {
log?.warn?.("TOKEN_REFRESH", "No Kimi refresh URL/clientId configured");
return null;
}
if (!refreshToken) return null;
return dedupRefresh("kimi", refreshToken, async () => {
try {
const headers = {
"Content-Type": "application/x-www-form-urlencoded",
Accept: "application/json",
...buildKimiHeaders(credentials?.providerSpecificData?.deviceId),
};
const response = await fetch(config.refreshUrl, {
method: "POST",
headers,
body: new URLSearchParams({
grant_type: "refresh_token",
refresh_token: refreshToken,
client_id: config.clientId,
}),
});
if (!response.ok) {
const errorText = await response.text();
log?.error?.("TOKEN_REFRESH", `Failed to refresh token for kimi`, {
status: response.status,
error: errorText,
});
return null;
}
const tokens = await response.json();
return {
accessToken: tokens.access_token,
refreshToken: tokens.refresh_token || refreshToken,
expiresIn: tokens.expires_in,
};
} catch (error) {
log?.error?.("TOKEN_REFRESH", `Error refreshing token for kimi`, { error: error.message });
return null;
}
}, log);
}
export async function refreshClaudeOAuthToken(refreshToken, log) {
if (!refreshToken) return null;
return dedupRefresh("claude", refreshToken, async () => {
open-sse/translator/request/claude-to-kiro.js
+6 -2
View File
@@ -33,6 +33,7 @@ import {
KIRO_AGENTIC_SYSTEM_PROMPT,
resolveDefaultProfileArn,
buildKiroAdditionalModelRequestFieldsForModel,
usesKiroNativeGptEffort,
} from "../../config/kiroConstants.js";
import { DEFAULT_IMAGE_MIME } from "../schema/index.js";
import { ROLE, CLAUDE_BLOCK } from "../schema/index.js";
@@ -390,6 +391,8 @@ export function claudeToKiroRequest(model, body, stream, credentials) {
const { upstream: upstreamModel, agentic } = resolveKiroModel(model);
const thinkingBudget = resolveKiroThinkingBudget(body, credentials?.rawHeaders, model);
const additionalModelRequestFields = buildKiroAdditionalModelRequestFieldsForModel(body, upstreamModel);
const usesNativeGptEffort = usesKiroNativeGptEffort(body, upstreamModel);
// Guard 1: no client tools → flatten all tool interactions to text.
if (!clientProvidedTools) {
@@ -421,7 +424,9 @@ export function claudeToKiroRequest(model, body, stream, credentials) {
// enforce top-level systemPrompt for direct calls.
const timestamp = new Date().toISOString();
const systemPromptParts = [];
if (thinkingBudget !== null) systemPromptParts.push(buildThinkingSystemPrefix(thinkingBudget));
if (thinkingBudget !== null && !usesNativeGptEffort) {
systemPromptParts.push(buildThinkingSystemPrefix(thinkingBudget));
}
if (agentic) systemPromptParts.push(KIRO_AGENTIC_SYSTEM_PROMPT);
const systemInstruction = extractClaudeSystemText(body.system);
if (systemInstruction) systemPromptParts.push(systemInstruction);
@@ -481,7 +486,6 @@ export function claudeToKiroRequest(model, body, stream, credentials) {
if (profileArn) payload.profileArn = profileArn;
if (systemPrompt) payload.systemPrompt = systemPrompt;
const additionalModelRequestFields = buildKiroAdditionalModelRequestFieldsForModel(body, upstreamModel);
if (additionalModelRequestFields) {
payload.additionalModelRequestFields = additionalModelRequestFields;
}
open-sse/translator/request/openai-responses.js
+4
View File
@@ -200,6 +200,9 @@ export function openaiResponsesToOpenAIRequest(model, body, stream, credentials)
delete result.include;
delete result.prompt_cache_key;
delete result.store;
if (typeof result.reasoning?.effort === "string") {
result.reasoning_effort = result.reasoning.effort;
}
delete result.reasoning;
delete result.client_metadata;
@@ -377,6 +380,7 @@ export function openaiToOpenAIResponsesRequest(model, body, stream, credentials)
if (body.top_p !== undefined) result.top_p = body.top_p;
if (body.reasoning !== undefined) result.reasoning = body.reasoning;
if (body.reasoning_effort !== undefined) result.reasoning = { effort: body.reasoning_effort, summary: "auto" };
if (body.service_tier !== undefined) result.service_tier = body.service_tier;
return result;
}
open-sse/translator/request/openai-to-kiro.js
+8 -8
View File
@@ -13,7 +13,8 @@ import {
buildThinkingSystemPrefix,
KIRO_AGENTIC_SYSTEM_PROMPT,
resolveDefaultProfileArn,
buildKiroAdditionalModelRequestFieldsForModel
buildKiroAdditionalModelRequestFieldsForModel,
usesKiroNativeGptEffort
} from "../../config/kiroConstants.js";
import { parseDataUri } from "../concerns/image.js";
import { DEFAULT_IMAGE_MIME } from "../schema/index.js";
@@ -511,12 +512,10 @@ function convertMessages(messages, tools, model) {
* Kiro's 2-3 minute server timeout. The suffix is stripped before being
* sent upstream.
*
* 2. Thinking / reasoning. Kiro does not accept `thinking.type` or
* `reasoning_effort` natively. The only way to enable reasoning is to
* inject `<thinking_mode>enabled</thinking_mode>` into the user content
* sent upstream. Detection covers Anthropic-Beta header, Claude API
* 2. Thinking / reasoning. Detection covers Anthropic-Beta header, Claude API
* `thinking`, OpenAI `reasoning_effort`, AMP/Cursor magic tags, and model
* name hints.
* name hints. Supported models receive Kiro's schema-specific effort fields;
* legacy prompt tags remain only for models that need them.
*/
export function openaiToKiroRequest(model, body, stream, credentials) {
const messages = body.messages || [];
@@ -527,6 +526,8 @@ export function openaiToKiroRequest(model, body, stream, credentials) {
const { upstream: upstreamModel, agentic } = resolveKiroModel(model);
const thinkingBudget = resolveKiroThinkingBudget(body, credentials?.rawHeaders, model);
const additionalModelRequestFields = buildKiroAdditionalModelRequestFieldsForModel(body, upstreamModel);
const usesNativeGptEffort = usesKiroNativeGptEffort(body, upstreamModel);
const { history, currentMessage } = convertMessages(messages, tools, upstreamModel);
@@ -554,7 +555,7 @@ export function openaiToKiroRequest(model, body, stream, credentials) {
// too because the CodeWhisperer surface does not always enforce top-level
// systemPrompt for direct calls.
const systemPromptParts = [];
if (thinkingBudget !== null) {
if (thinkingBudget !== null && !usesNativeGptEffort) {
systemPromptParts.push(buildThinkingSystemPrefix(thinkingBudget));
}
if (agentic) {
@@ -612,7 +613,6 @@ export function openaiToKiroRequest(model, body, stream, credentials) {
payload.profileArn = profileArn;
}
if (systemPrompt) payload.systemPrompt = systemPrompt;
const additionalModelRequestFields = buildKiroAdditionalModelRequestFieldsForModel(body, upstreamModel);
if (additionalModelRequestFields) {
payload.additionalModelRequestFields = additionalModelRequestFields;
}
open-sse/utils/cursorChecksum.js
+2 -1
View File
@@ -128,7 +128,8 @@ export function buildCursorHeaders(accessToken, machineId = null, ghostMode = tr
"x-amzn-trace-id": `Root=${crypto.randomUUID()}`,
"x-client-key": clientKey,
"x-cursor-checksum": checksum,
"x-cursor-client-version": "3.1.0",
"x-cursor-client-version": "3.12.17",
"x-cursor-client-commit": "0fb762053c34788bb7760d5673f8a6d4c8589d50",
"x-cursor-client-type": "ide",
"x-cursor-client-os": os,
"x-cursor-client-arch": arch,
package.json
+1 -1
View File
@@ -1,6 +1,6 @@
{
"name": "9router-app",
"version": "0.5.35",
"version": "0.5.40",
"description": "9Router web dashboard",
"private": true,
"scripts": {
public/i18n/literals/km.json
+1396
View File
File diff suppressed because it is too large Load Diff
public/providers/agentrouter.png
BIN
View File
Binary file not shown.
Side by Side

After

Width:  |  Height:  |  Size: 7.7 KiB

public/providers/codebuddy-cn.png
BIN
View File
Binary file not shown.
Side by Side Swipe Overlay

Before

Width:  |  Height:  |  Size: 23 KiB

After

Width:  |  Height:  |  Size: 17 KiB

public/providers/commandcode.png
BIN
View File
Binary file not shown.
Side by Side Swipe Overlay

Before

Width:  |  Height:  |  Size: 6.4 KiB

After

Width:  |  Height:  |  Size: 9.1 KiB

public/providers/featherless.png
BIN
View File
Binary file not shown.
Side by Side

After

Width:  |  Height:  |  Size: 3.8 KiB

public/providers/gitlab.png
BIN
View File
Binary file not shown.
Side by Side

After

Width:  |  Height:  |  Size: 3.0 KiB

public/providers/longcat.png
BIN
View File
Binary file not shown.
Side by Side

After

Width:  |  Height:  |  Size: 5.8 KiB

public/providers/mmf.png
BIN
View File
Binary file not shown.
Side by Side

After

Width:  |  Height:  |  Size: 2.6 KiB

public/providers/morph.png
BIN
View File
Binary file not shown.
Side by Side

After

Width:  |  Height:  |  Size: 2.5 KiB

public/providers/novita.png
BIN
View File
Binary file not shown.
Side by Side

After

Width:  |  Height:  |  Size: 1.3 KiB

public/providers/perplexity-agent.png
BIN
View File
Binary file not shown.
Side by Side

After

Width:  |  Height:  |  Size: 10 KiB

public/providers/qoder.png
BIN
View File
Binary file not shown.
Side by Side Swipe Overlay

Before

Width:  |  Height:  |  Size: 7.0 KiB

After

Width:  |  Height:  |  Size: 5.9 KiB

public/providers/reka.png
BIN
View File
Binary file not shown.
Side by Side

After

Width:  |  Height:  |  Size: 2.4 KiB

public/providers/venice.png
BIN
View File
Binary file not shown.
Side by Side

After

Width:  |  Height:  |  Size: 5.9 KiB

public/providers/vercel-ai-gateway.png
BIN
View File
Binary file not shown.
Side by Side

After

Width:  |  Height:  |  Size: 2.0 KiB

public/providers/vercel.png
BIN
View File
Binary file not shown.
Side by Side

After

Width:  |  Height:  |  Size: 2.0 KiB

src/app/(dashboard)/dashboard/basic-chat/BasicChatPageClient.js
+1 -1
View File
@@ -792,7 +792,7 @@ export default function BasicChatPageClient() {
<div className="mb-3 grid grid-cols-2 gap-2 sm:grid-cols-3 mt-2">
{message.attachments.map((attachment) => (
<a key={attachment.id} href={attachment.dataUrl} target="_blank" rel="noreferrer" className="overflow-hidden rounded-[18px] border border-white/10 bg-black/20">
<img src={attachment.dataUrl} alt={attachment.name} className="h-28 w-full object-cover" />
<img src={attachment.dataUrl} alt={attachment.name} className="h-28 w-full object-cover" loading="lazy" decoding="async" />
</a>
))}
</div>
src/app/(dashboard)/dashboard/cli-tools/components/AntigravityToolCard.js
+17 -18
View File
@@ -38,15 +38,10 @@ export default function AntigravityToolCard({
}, [initialStatus]);
useEffect(() => {
if (isExpanded && !status) {
fetchStatus();
loadSavedMappings();
fetchModelAliases();
}
if (isExpanded) {
loadSavedMappings();
fetchModelAliases();
}
if (!isExpanded) return;
if (!status) fetchStatus();
loadSavedMappings();
fetchModelAliases();
}, [isExpanded]);
const loadSavedMappings = async () => {
@@ -243,6 +238,8 @@ export default function AntigravityToolCard({
className="size-8 object-contain rounded-lg"
sizes="32px"
onError={(e) => { e.target.style.display = "none"; }}
loading="lazy"
decoding="async"
/>
</div>
<div className="min-w-0">
@@ -467,15 +464,17 @@ export default function AntigravityToolCard({
</Modal>
{/* Model Select Modal */}
<ModelSelectModal
isOpen={modalOpen}
onClose={() => setModalOpen(false)}
onSelect={handleModelSelect}
selectedModel={currentEditingAlias ? modelMappings[currentEditingAlias] : null}
activeProviders={activeProviders}
modelAliases={modelAliases}
title={`Select model for ${currentEditingAlias}`}
/>
{modalOpen && (
<ModelSelectModal
isOpen={modalOpen}
onClose={() => setModalOpen(false)}
onSelect={handleModelSelect}
selectedModel={currentEditingAlias ? modelMappings[currentEditingAlias] : null}
activeProviders={activeProviders}
modelAliases={modelAliases}
title={`Select model for ${currentEditingAlias}`}
/>
)}
</Card>
);
}
src/app/(dashboard)/dashboard/cli-tools/components/DefaultToolCard.js
+28 -14
View File
@@ -2,6 +2,7 @@
import { useEffect, useRef, useState } from "react";
import { Button, Card, ModelSelectModal } from "@/shared/components";
import { getProviderIconSrc, markProviderIconMissing } from "@/shared/utils/providerIcon";
import { useCopyToClipboard } from "@/shared/hooks/useCopyToClipboard";
import Image from "next/image";
import ApiKeySelect from "./ApiKeySelect";
@@ -332,21 +333,32 @@ export default function DefaultToolCard({ toolId, tool, baseUrl, apiKeys, active
className="size-8 object-contain rounded-lg"
sizes="32px"
onError={(e) => { e.target.style.display = "none"; }}
loading="lazy"
decoding="async"
/>
);
}
if (tool.icon) {
return <span className="material-symbols-outlined text-xl" style={{ color: tool.color }}>{tool.icon}</span>;
}
const iconSrc = getProviderIconSrc(toolId);
if (!iconSrc) {
return <span className="text-xs font-bold" style={{ color: tool.color }}>{(toolId || "?").slice(0, 2).toUpperCase()}</span>;
}
return (
<Image
src={`/providers/${toolId}.png`}
src={iconSrc}
alt={tool.name}
width={32}
height={32}
className="size-8 object-contain rounded-lg"
sizes="32px"
onError={(e) => { e.target.style.display = "none"; }}
onError={(e) => {
markProviderIconMissing(toolId);
e.target.style.display = "none";
}}
loading="lazy"
decoding="async"
/>
);
};
@@ -377,18 +389,20 @@ export default function DefaultToolCard({ toolId, tool, baseUrl, apiKeys, active
</div>
)}
<ModelSelectModal
isOpen={showModelModal}
onClose={() => setShowModelModal(false)}
onSelect={tool.modelSelection === "multiple" ? addSelectedModel : handleSelectModel}
selectedModel={modelValue}
activeProviders={activeProviders}
title={tool.modelSelection === "multiple" ? "Add Cursor custom model" : "Select Model"}
closeOnSelect={tool.modelSelection !== "multiple"}
addedModelValues={tool.modelSelection === "multiple" ? selectedModels : []}
allowDuplicates={tool.modelSelection === "multiple"}
availableModels={availableModels}
/>
{showModelModal && (
<ModelSelectModal
isOpen={showModelModal}
onClose={() => setShowModelModal(false)}
onSelect={tool.modelSelection === "multiple" ? addSelectedModel : handleSelectModel}
selectedModel={modelValue}
activeProviders={activeProviders}
title={tool.modelSelection === "multiple" ? "Add Cursor custom model" : "Select Model"}
closeOnSelect={tool.modelSelection !== "multiple"}
addedModelValues={tool.modelSelection === "multiple" ? selectedModels : []}
allowDuplicates={tool.modelSelection === "multiple"}
availableModels={availableModels}
/>
)}
</Card>
);
}
src/app/(dashboard)/dashboard/cli-tools/components/MitmLinkCard.js
+2
View File
@@ -22,6 +22,8 @@ export default function MitmLinkCard({ tool }) {
className="size-8 object-contain rounded-lg"
sizes="32px"
onError={(e) => { e.target.style.display = "none"; }}
loading="lazy"
decoding="async"
/>
</div>
<div className="min-w-0">
src/app/(dashboard)/dashboard/cli-tools/components/MitmToolCard.js
+14 -10
View File
@@ -146,6 +146,8 @@ export default function MitmToolCard({
className="size-8 object-contain rounded-lg"
sizes="32px"
onError={(e) => { e.target.style.display = "none"; }}
loading="lazy"
decoding="async"
/>
</div>
<div className="min-w-0">
@@ -307,16 +309,18 @@ export default function MitmToolCard({
)}
{/* Model Select Modal */}
<ModelSelectModal
isOpen={modalOpen}
onClose={() => setModalOpen(false)}
onSelect={handleModelSelect}
selectedModel={currentEditingAlias ? modelMappings[currentEditingAlias] : null}
activeProviders={activeProviders}
modelAliases={modelAliases}
availableModels={availableModels}
title={`Select model for ${currentEditingAlias}`}
/>
{modalOpen && (
<ModelSelectModal
isOpen={modalOpen}
onClose={() => setModalOpen(false)}
onSelect={handleModelSelect}
selectedModel={currentEditingAlias ? modelMappings[currentEditingAlias] : null}
activeProviders={activeProviders}
modelAliases={modelAliases}
availableModels={availableModels}
title={`Select model for ${currentEditingAlias}`}
/>
)}
</>
);
}
src/app/(dashboard)/dashboard/cli-tools/components/ToolSummaryCard.js
+1 -1
View File
@@ -12,7 +12,7 @@ export default function ToolSummaryCard({ toolId, tool }) {
<div className="flex items-center gap-3">
<div className="size-8 flex items-center justify-center shrink-0">
{tool.image ? (
<Image src={tool.image} alt={tool.name} width={32} height={32} className="size-8 object-contain rounded-lg" sizes="32px" onError={(e) => { e.target.style.display = "none"; }} />
<Image src={tool.image} alt={tool.name} width={32} height={32} className="size-8 object-contain rounded-lg" sizes="32px" onError={(e) => { e.target.style.display = "none"; }} loading="lazy" decoding="async" />
) : tool.icon ? (
<span className="material-symbols-outlined text-[28px]" style={{ color: tool.color }}>{tool.icon}</span>
) : null}
src/app/(dashboard)/dashboard/combos/page.js
+44 -36
View File
@@ -24,7 +24,7 @@ export default function CombosPage() {
useEffect(() => {
fetchData();
}, []); // eslint-disable-line react-hooks/exhaustive-deps
}, []);
async function fetchData() {
try {
@@ -205,23 +205,27 @@ export default function CombosPage() {
)}
{/* Create Modal - Use key to force remount and reset state */}
<ComboFormModal
key="create"
isOpen={showCreateModal}
onClose={() => setShowCreateModal(false)}
onSave={handleCreate}
availableModels={selectableModels}
/>
{showCreateModal && (
<ComboFormModal
key="create"
isOpen={showCreateModal}
onClose={() => setShowCreateModal(false)}
onSave={handleCreate}
availableModels={selectableModels}
/>
)}
{/* Edit Modal - Use key to force remount and reset state */}
<ComboFormModal
key={editingCombo?.id || "new"}
isOpen={!!editingCombo}
combo={editingCombo}
onClose={() => setEditingCombo(null)}
onSave={(data) => handleUpdate(editingCombo.id, data)}
availableModels={selectableModels}
/>
{editingCombo && (
<ComboFormModal
key={editingCombo.id}
isOpen={true}
combo={editingCombo}
onClose={() => setEditingCombo(null)}
onSave={(data) => handleUpdate(editingCombo.id, data)}
availableModels={selectableModels}
/>
)}
{/* Confirm Delete Modal */}
<ConfirmModal
@@ -346,15 +350,17 @@ function ComboCard({ combo, modelCaps = {}, availableModels = [], copied, onCopy
</div>
{/* Judge model picker (single-select; combo members make natural judges too) */}
<ModelSelectModal
isOpen={showJudgeSelect}
onClose={() => setShowJudgeSelect(false)}
onSelect={(m) => { onSetStrategy({ judgeModel: m?.value || "" }); setShowJudgeSelect(false); }}
availableModels={availableModels}
title="Select Judge Model"
addedModelValues={judge ? [judge] : []}
closeOnSelect={true}
/>
{showJudgeSelect && (
<ModelSelectModal
isOpen={showJudgeSelect}
onClose={() => setShowJudgeSelect(false)}
onSelect={(m) => { onSetStrategy({ judgeModel: m?.value || "" }); setShowJudgeSelect(false); }}
availableModels={availableModels}
title="Select Judge Model"
addedModelValues={judge ? [judge] : []}
closeOnSelect={true}
/>
)}
</Card>
);
}
@@ -631,17 +637,19 @@ function ComboFormModal({ isOpen, combo, onClose, onSave, availableModels = [],
</Modal>
{/* Model Select Modal */}
<ModelSelectModal
isOpen={showModelSelect}
onClose={() => setShowModelSelect(false)}
onSelect={handleAddModel}
onDeselect={handleDeselectModel}
availableModels={availableModels}
title="Add Model to Combo"
kindFilter={kindFilter}
addedModelValues={models}
closeOnSelect={false}
/>
{showModelSelect && (
<ModelSelectModal
isOpen={showModelSelect}
onClose={() => setShowModelSelect(false)}
onSelect={handleAddModel}
onDeselect={handleDeselectModel}
availableModels={availableModels}
title="Add Model to Combo"
kindFilter={kindFilter}
addedModelValues={models}
closeOnSelect={false}
/>
)}
</>
);
}
src/app/(dashboard)/dashboard/media-providers/[kind]/[id]/components/GenericExampleCard.js
+8
View File
@@ -332,6 +332,8 @@ export function GenericExampleCard({ providerId, kind }) {
className="max-h-40 rounded-lg border border-border object-contain bg-sidebar"
onError={(e) => { e.currentTarget.style.display = "none"; }}
onLoad={(e) => { e.currentTarget.style.display = "block"; }}
loading="lazy"
decoding="async"
/>
)}
</div>
@@ -365,6 +367,8 @@ export function GenericExampleCard({ providerId, kind }) {
className="max-h-40 rounded-lg border border-border object-contain bg-sidebar"
onError={(e) => { e.currentTarget.style.display = "none"; }}
onLoad={(e) => { e.currentTarget.style.display = "block"; }}
loading="lazy"
decoding="async"
/>
)}
</div>
@@ -469,6 +473,8 @@ export function GenericExampleCard({ providerId, kind }) {
src={`data:image/png;base64,${partialImage.b64_json}`}
alt="Partial"
className="max-w-full rounded-lg border border-border mt-1.5 opacity-80"
loading="lazy"
decoding="async"
/>
</div>
)}
@@ -511,6 +517,8 @@ export function GenericExampleCard({ providerId, kind }) {
src={binaryImageUrl || (result?.data?.data?.[0]?.b64_json ? `data:image/png;base64,${result.data.data[0].b64_json}` : result?.data?.data?.[0]?.url)}
alt="Generated"
className="max-w-full rounded-lg border border-border"
loading="lazy"
decoding="async"
/>
</div>
)}
src/app/(dashboard)/dashboard/media-providers/combo/[id]/page.js
+15 -13
View File
@@ -352,7 +352,7 @@ export default function ComboDetailPage() {
Download
</a>
</div>
<img src={testResult.imageUrl} alt="Generated" className="max-w-full rounded-lg border border-border" />
<img src={testResult.imageUrl} alt="Generated" className="max-w-full rounded-lg border border-border" loading="lazy" decoding="async" />
</div>
)}
{testResult.audioUrl && (
@@ -388,18 +388,20 @@ export default function ComboDetailPage() {
)}
</Card>
<ModelSelectModal
isOpen={showPicker}
onClose={() => setShowPicker(false)}
onSelect={handleAddModel}
onDeselect={handleDeselectModel}
activeProviders={connections}
modelAliases={modelAliases}
title={`Add ${kindLabel} Model`}
kindFilter={combo.kind}
addedModelValues={providers}
closeOnSelect={false}
/>
{showPicker && (
<ModelSelectModal
isOpen={showPicker}
onClose={() => setShowPicker(false)}
onSelect={handleAddModel}
onDeselect={handleDeselectModel}
activeProviders={connections}
modelAliases={modelAliases}
title={`Add ${kindLabel} Model`}
kindFilter={combo.kind}
addedModelValues={providers}
closeOnSelect={false}
/>
)}
</div>
);
}
src/app/(dashboard)/dashboard/providers/[id]/page.js
+47 -5
View File
@@ -4,6 +4,7 @@ import { useState, useEffect, useCallback, useRef } from "react";
import { useParams, useRouter } from "next/navigation";
import Link from "next/link";
import Image from "next/image";
import { getProviderIconSrc, markProviderIconMissing } from "@/shared/utils/providerIcon";
import { Card, Button, Badge, Input, Modal, CardSkeleton, OAuthModal, KiroOAuthWrapper, CursorAuthModal, IFlowCookieModal, GitLabAuthModal, Toggle, Select, EditConnectionModal, NoAuthProxyCard, ConfirmModal } from "@/shared/components";
import { OAUTH_PROVIDERS, APIKEY_PROVIDERS, FREE_PROVIDERS, FREE_TIER_PROVIDERS, WEB_COOKIE_PROVIDERS, getProviderAlias, isOpenAICompatibleProvider, isAnthropicCompatibleProvider, AI_PROVIDERS } from "@/shared/constants/providers";
import { getModelsByProviderId, getModelKind } from "@/shared/constants/models";
@@ -66,6 +67,7 @@ export default function ProviderDetailPage() {
const [thinkingMode, setThinkingMode] = useState("auto");
const [autoPing, setAutoPing] = useState({ enabled: false, connections: {} });
const [suggestedModels, setSuggestedModels] = useState([]);
const [liveModels, setLiveModels] = useState([]);
const [kiloFreeModels, setKiloFreeModels] = useState([]);
const [deletedModelIds, setDeletedModelIds] = useState([]);
const [confirmState, setConfirmState] = useState(null);
@@ -141,7 +143,10 @@ export default function ProviderDetailPage() {
const isOAuth = !!OAUTH_PROVIDERS[providerId] || !!FREE_PROVIDERS[providerId] || authModes.includes("oauth");
const supportsApiKeyAuth = !!APIKEY_PROVIDERS[providerId] || authModes.includes("apikey");
const isFreeNoAuth = !!FREE_PROVIDERS[providerId]?.noAuth;
const models = getModelsByProviderId(providerId);
const staticModels = getModelsByProviderId(providerId);
const models = providerId === "cursor" && liveModels.length > 0
? liveModels
: staticModels;
const providerAlias = getProviderAlias(providerId);
const isOpenAICompatible = isOpenAICompatibleProvider(providerId);
@@ -151,8 +156,12 @@ export default function ProviderDetailPage() {
const oauthConnectionLabel =
providerId === "xai" ? "Grok Build OAuth"
: providerId === "grok-cli" ? "Grok CLI Device Login"
: providerId === "kimi" ? "Kimi Coding OAuth"
: "OAuth";
const apiKeyConnectionLabel = providerId === "xai" ? "xAI API Key" : "API Key";
const apiKeyConnectionLabel =
providerId === "xai" ? "xAI API Key"
: providerId === "kimi" ? "Kimi API Key"
: "API Key";
// Resolve suffix "(level)" for a model when a thinking level is picked and the model supports it.
const resolveThinkingSuffix = (modelId) => {
if (!thinkingMode || thinkingMode === "auto") return null;
@@ -427,6 +436,34 @@ export default function ProviderDetailPage() {
fetchDeletedModels();
}, [fetchConnections, fetchAliases, fetchCustomModels, fetchDeletedModels]);
// Cursor's model availability is account-specific and changes frequently.
// Load the active account's live catalog for the dashboard; the static
// registry remains the fallback while the request is pending or unavailable.
useEffect(() => {
if (providerId !== "cursor") {
setLiveModels([]);
return;
}
const connection = connections.find((item) => item.isActive !== false);
if (!connection?.id) {
setLiveModels([]);
return;
}
let cancelled = false;
fetch(`/api/providers/${connection.id}/models`, { cache: "no-store" })
.then(async (res) => ({ ok: res.ok, data: await res.json() }))
.then(({ ok, data }) => {
if (!cancelled && ok && Array.isArray(data.models) && data.models.length > 0) {
setLiveModels(data.models);
}
})
.catch(() => {});
return () => { cancelled = true; };
}, [providerId, connections]);
// Fetch suggested models from provider's public API (if configured)
useEffect(() => {
const fetcher = (OAUTH_PROVIDERS[providerId] || APIKEY_PROVIDERS[providerId] || FREE_PROVIDERS[providerId] || FREE_TIER_PROVIDERS[providerId])?.modelsFetcher;
@@ -1196,7 +1233,7 @@ export default function ProviderDetailPage() {
if (isAnthropicCompatible) {
return "/providers/anthropic-m.png";
}
return `/providers/${providerInfo.id}.png`;
return getProviderIconSrc(providerInfo.id);
};
return (
@@ -1215,7 +1252,7 @@ export default function ProviderDetailPage() {
className="flex size-12 shrink-0 items-center justify-center rounded-lg"
style={{ backgroundColor: `${providerInfo.color}15` }}
>
{headerImgError ? (
{headerImgError || !getHeaderIconPath() ? (
<span className="text-sm font-bold" style={{ color: providerInfo.color }}>
{providerInfo.textIcon || providerInfo.id.slice(0, 2).toUpperCase()}
</span>
@@ -1227,7 +1264,12 @@ export default function ProviderDetailPage() {
height={48}
className="max-h-12 max-w-12 rounded-lg object-contain"
sizes="48px"
onError={() => setHeaderImgError(true)}
onError={() => {
markProviderIconMissing(providerInfo.id);
setHeaderImgError(true);
}}
loading="lazy"
decoding="async"
/>
)}
</div>
src/app/(dashboard)/dashboard/providers/components/ModelsCard.js
+4 -8
View File
@@ -116,26 +116,22 @@ export default function ModelsCard({ providerId, kindFilter, providerAliasOverri
const [testingModelId, setTestingModelId] = useState(null);
const [testError, setTestError] = useState("");
const [showAddCustomModel, setShowAddCustomModel] = useState(false);
const [connections, setConnections] = useState([]);
const providerAlias = providerAliasOverride || getProviderAlias(providerId);
const effectiveType = kindFilter || "llm";
const fetchData = useCallback(async () => {
try {
const [aliasRes, connRes, customRes] = await Promise.all([
const [aliasRes, customRes] = await Promise.all([
fetch("/api/models/alias"),
fetch("/api/providers", { cache: "no-store" }),
fetch("/api/models/custom", { cache: "no-store" }),
]);
const aliasData = await aliasRes.json();
const connData = await connRes.json();
const customData = await customRes.json();
if (aliasRes.ok) setModelAliases(aliasData.aliases || {});
if (connRes.ok) setConnections((connData.connections || []).filter((c) => c.provider === providerId));
if (customRes.ok) setCustomModels(customData.models || []);
} catch (e) { console.log("ModelsCard fetch error:", e); }
}, [providerId]);
}, []);
useEffect(() => { fetchData(); }, [fetchData]);
@@ -242,7 +238,7 @@ export default function ModelsCard({ providerId, kindFilter, providerAliasOverri
onSetAlias={(alias) => handleSetAlias(model.id, alias)}
onDeleteAlias={() => handleDeleteAlias(existingAlias)}
testStatus={modelTestResults[model.id]}
onTest={connections.length > 0 ? () => handleTestModel(model.id) : undefined}
onTest={() => handleTestModel(model.id)}
isTesting={testingModelId === model.id}
isFree={model.isFree}
/>
@@ -259,7 +255,7 @@ export default function ModelsCard({ providerId, kindFilter, providerAliasOverri
onSetAlias={() => {}}
onDeleteAlias={() => handleDeleteCustomModel(model.id)}
testStatus={modelTestResults[model.id]}
onTest={connections.length > 0 ? () => handleTestModel(model.id) : undefined}
onTest={() => handleTestModel(model.id)}
isTesting={testingModelId === model.id}
isCustom
/>
src/app/(dashboard)/dashboard/providers/page.js
+3 -2
View File
@@ -10,6 +10,7 @@ import {
Toggle,
} from "@/shared/components";
import ProviderIcon from "@/shared/components/ProviderIcon";
import { getProviderIconSrc } from "@/shared/utils/providerIcon";
import { OAUTH_PROVIDERS, APIKEY_PROVIDERS } from "@/shared/constants/config";
import {
FREE_PROVIDERS,
@@ -761,12 +762,12 @@ function ApiKeyProviderCard({
};
const getIconPath = () => {
if (isCompatible)
if (isCompatible && provider.apiType)
return provider.apiType === "responses"
? "/providers/oai-r.png"
: "/providers/oai-cc.png";
if (isAnthropicCompatible) return "/providers/anthropic-m.png";
return `/providers/${provider.id}.png`;
return getProviderIconSrc(provider.id);
};
return (
src/app/(dashboard)/dashboard/usage/components/ProviderLimits/index.js
+16 -14
View File
@@ -1236,22 +1236,24 @@ export default function ProviderLimits() {
/>
)}
{hiddenQuotaRows.length > 0 && (
<div className="mt-2 flex flex-wrap items-center gap-1 border-t border-black/5 pt-2 text-[10px] text-text-muted dark:border-white/5">
<span className="material-symbols-outlined text-[14px]">
<div className="mt-2 flex min-w-0 items-center gap-1 border-t border-black/5 pt-2 text-[10px] text-text-muted dark:border-white/5">
<span className="material-symbols-outlined shrink-0 text-[14px]">
visibility_off
</span>
<span>Hidden:</span>
{hiddenQuotaRows.map((quotaRow) => (
<button
key={getQuotaVisibilityKey(quotaRow)}
type="button"
onClick={() => handleShowQuota(conn.provider, quotaRow)}
className="rounded-md border border-black/10 px-1.5 py-0.5 transition-colors hover:bg-black/5 hover:text-text-primary dark:border-white/10 dark:hover:bg-white/5"
title="Show this quota row"
>
{quotaRow.name}
</button>
))}
<span className="shrink-0">Hidden:</span>
<div className="flex min-w-0 flex-1 items-center gap-1 overflow-x-auto whitespace-nowrap">
{hiddenQuotaRows.map((quotaRow) => (
<button
key={getQuotaVisibilityKey(quotaRow)}
type="button"
onClick={() => handleShowQuota(conn.provider, quotaRow)}
className="shrink-0 rounded-md border border-black/10 px-1.5 py-0.5 transition-colors hover:bg-black/5 hover:text-text-primary dark:border-white/10 dark:hover:bg-white/5"
title="Show this quota row"
>
{quotaRow.name}
</button>
))}
</div>
</div>
)}
</div>
src/app/(dashboard)/dashboard/usage/components/ProviderTopology.js
+174 -14
View File
@@ -7,21 +7,27 @@ import {
Handle,
Position,
Controls,
BaseEdge,
getBezierPath,
} from "@xyflow/react";
import "@xyflow/react/dist/style.css";
import { AI_PROVIDERS } from "@/shared/constants/providers";
import { getProviderIconSrc, markProviderIconMissing } from "@/shared/utils/providerIcon";
// Force-stop FE animation if a provider stays active longer than this
const FE_ACTIVE_TIMEOUT_MS = 60000;
const FE_ACTIVE_TICK_MS = 1000;
// Kame + electric particles along active edges
const KAME_PARTICLE_COUNT = 6;
const SPARK_COUNT = 5;
function getProviderConfig(providerId) {
return AI_PROVIDERS[providerId] || { color: "#6b7280", name: providerId };
}
// Use local provider images from /public/providers/
function getProviderImageUrl(providerId) {
return `/providers/${providerId}.png`;
return getProviderIconSrc(providerId);
}
// Custom provider node - rectangle with image + name
@@ -47,8 +53,19 @@ function ProviderNode({ data }) {
className="w-8 h-8 rounded-md flex items-center justify-center shrink-0"
style={{ backgroundColor: `${color}15` }}
>
{!imgError ? (
<img src={imageUrl} alt={label} className="w-6 h-6 rounded-sm object-contain" onError={() => setImgError(true)} />
{imageUrl && !imgError ? (
<img
src={imageUrl}
alt={label}
className="w-6 h-6 rounded-sm object-contain"
loading="lazy"
decoding="async"
onError={() => {
const m = imageUrl?.match(/^\/providers\/([^/]+)\.png$/i);
if (m) markProviderIconMissing(m[1]);
setImgError(true);
}}
/>
) : (
<span className="text-sm font-bold" style={{ color }}>{textIcon}</span>
)}
@@ -77,19 +94,34 @@ ProviderNode.propTypes = {
data: PropTypes.object.isRequired,
};
// Center 9Router node
// Center 9Router node — pulse/glow on card only (no expanding rings)
function RouterNode({ data }) {
const powering = (data.activeCount || 0) > 0;
return (
<div className="flex items-center justify-center px-5 py-3 rounded-xl border-2 border-primary bg-primary/5 shadow-md min-w-[130px]">
<div
className={`relative z-[1] flex items-center justify-center px-5 py-3 rounded-xl border-2 min-w-[130px] ${
powering
? "topology-router-core border-yellow-300 bg-gradient-to-br from-primary/30 via-yellow-400/20 to-cyan-400/25"
: "border-primary bg-primary/5 shadow-md"
}`}
>
<Handle type="source" position={Position.Top} id="top" className="!bg-transparent !border-0 !w-0 !h-0" />
<Handle type="source" position={Position.Bottom} id="bottom" className="!bg-transparent !border-0 !w-0 !h-0" />
<Handle type="source" position={Position.Left} id="left" className="!bg-transparent !border-0 !w-0 !h-0" />
<Handle type="source" position={Position.Right} id="right" className="!bg-transparent !border-0 !w-0 !h-0" />
<img src="/favicon.svg" alt="9Router (Remake)" className="w-6 h-6 mr-2" />
<span className="text-sm font-bold text-primary">9Router (Remake)</span>
<img
src="/favicon.svg"
alt="9Router"
className={`w-6 h-6 mr-2 ${powering ? "topology-router-icon" : ""}`}
loading="lazy"
decoding="async"
/>
<span className={`text-sm font-bold ${powering ? "topology-router-label text-yellow-300" : "text-primary"}`}>
9Router
</span>
{data.activeCount > 0 && (
<span className="ml-2 px-1.5 py-0.5 rounded-full bg-primary text-white text-xs font-bold">
<span className="ml-2 px-1.5 py-0.5 rounded-full bg-yellow-400 text-black text-xs font-bold topology-router-badge">
{data.activeCount}
</span>
)}
@@ -101,7 +133,131 @@ RouterNode.propTypes = {
data: PropTypes.object.isRequired,
};
// Active: electric kame beam (multi-layer stroke + sparks). Idle/last/error: solid BaseEdge.
function TopologyEdge({
id,
sourceX,
sourceY,
targetX,
targetY,
sourcePosition,
targetPosition,
style = {},
data,
}) {
const [edgePath] = getBezierPath({
sourceX,
sourceY,
sourcePosition,
targetX,
targetY,
targetPosition,
});
const active = !!data?.active;
const stroke = style.stroke || "var(--color-border)";
const filterId = `topo-electric-${id}`;
if (!active) {
return <BaseEdge id={id} path={edgePath} style={{ ...style, stroke }} />;
}
return (
<g className="topology-edge-electric">
<defs>
<filter id={filterId} x="-40%" y="-40%" width="180%" height="180%">
<feTurbulence type="fractalNoise" baseFrequency="0.9" numOctaves="2" seed="2" result="noise">
<animate attributeName="baseFrequency" values="0.8;1.4;0.8" dur="0.25s" repeatCount="indefinite" />
</feTurbulence>
<feDisplacementMap in="SourceGraphic" in2="noise" scale="3.5" xChannelSelector="R" yChannelSelector="G" />
</filter>
</defs>
{/* Outer electric halo */}
<path
d={edgePath}
fill="none"
stroke="#22d3ee"
strokeWidth={10}
strokeOpacity={0.35}
strokeLinecap="round"
filter={`url(#${filterId})`}
className="topology-edge-halo"
/>
{/* Mid plasma */}
<path
d={edgePath}
fill="none"
stroke="#4ade80"
strokeWidth={5}
strokeOpacity={0.85}
strokeLinecap="round"
filter={`url(#${filterId})`}
className="topology-edge-plasma"
/>
{/* Hot white core */}
<BaseEdge
id={id}
path={edgePath}
style={{ stroke: "#f8fafc", strokeWidth: 2.2, opacity: 1 }}
className="topology-edge-kame"
/>
{/* Energy orbs */}
{Array.from({ length: KAME_PARTICLE_COUNT }, (_, i) => (
<circle
key={`${id}-p-${i}`}
r={i % 2 === 0 ? 4 : 2.5}
fill={i % 3 === 0 ? "#fde047" : i % 3 === 1 ? "#67e8f9" : "#fff"}
opacity={0.95}
style={{ filter: "drop-shadow(0 0 4px #22d3ee)" }}
>
<animateMotion
dur={`${0.4 + i * 0.08}s`}
repeatCount="indefinite"
path={edgePath}
begin={`${i * 0.09}s`}
/>
</circle>
))}
{/* Electric sparks (short-lived blink along path) */}
{Array.from({ length: SPARK_COUNT }, (_, i) => (
<circle
key={`${id}-s-${i}`}
r={1.8}
fill="#e0f2fe"
opacity={0}
>
<animate
attributeName="opacity"
values="0;1;0;0;1;0"
dur={`${0.35 + (i % 3) * 0.1}s`}
begin={`${i * 0.07}s`}
repeatCount="indefinite"
/>
<animateMotion
dur={`${0.28 + i * 0.05}s`}
repeatCount="indefinite"
path={edgePath}
begin={`${i * 0.11}s`}
/>
</circle>
))}
</g>
);
}
TopologyEdge.propTypes = {
id: PropTypes.string,
sourceX: PropTypes.number,
sourceY: PropTypes.number,
targetX: PropTypes.number,
targetY: PropTypes.number,
sourcePosition: PropTypes.string,
targetPosition: PropTypes.string,
style: PropTypes.object,
data: PropTypes.object,
};
const nodeTypes = { provider: ProviderNode, router: RouterNode };
const edgeTypes = { topology: TopologyEdge };
// Place N nodes evenly along an ellipse around the router center.
function buildLayout(providers, activeSet, lastSet, errorSet) {
@@ -135,9 +291,9 @@ function buildLayout(providers, activeSet, lastSet, errorSet) {
draggable: false,
});
const edgeStyle = (active, last, error, color) => {
const edgeStyle = (active, last, error) => {
if (error) return { stroke: "#ef4444", strokeWidth: 2.5, opacity: 0.9 };
if (active) return { stroke: "#22c55e", strokeWidth: 2.5, opacity: 0.9 };
if (active) return { stroke: "#22d3ee", strokeWidth: 3.5, opacity: 1 };
if (last) return { stroke: "#f59e0b", strokeWidth: 2, opacity: 0.7 };
return { stroke: "var(--color-border)", strokeWidth: 1, opacity: 0.3 };
};
@@ -183,12 +339,15 @@ function buildLayout(providers, activeSet, lastSet, errorSet) {
edges.push({
id: `e-${nodeId}`,
type: "topology",
source: "router",
sourceHandle,
target: nodeId,
targetHandle,
animated: active,
style: edgeStyle(active, last, error, config.color),
// Built-in animated uses stroke-dasharray (CPU-heavy); use particle beam instead
animated: false,
data: { active },
style: edgeStyle(active, last, error),
});
});
@@ -241,7 +400,7 @@ export default function ProviderTopology({ providers = [], activeRequests = [],
const { nodes, edges } = useMemo(
() => buildLayout(providers, activeSet, lastSet, errorSet),
[providers, activeSet, lastKey, errorKey]
[providers, activeSet, lastSet, errorSet]
);
// Stable key — only remount when provider list changes
@@ -289,6 +448,7 @@ export default function ProviderTopology({ providers = [], activeRequests = [],
nodes={nodes}
edges={edges}
nodeTypes={nodeTypes}
edgeTypes={edgeTypes}
fitView
fitViewOptions={fitOpts}
minZoom={0.1}
src/app/api/models/route.js
+10 -1
View File
@@ -21,12 +21,21 @@ export async function GET() {
})
.map((m) => {
const fullModel = `${m.provider}/${m.model}`;
const providerAlias = getProviderAlias(m.provider) || m.provider;
const routedModel = `${providerAlias}/${m.model}`;
const c = getCapabilitiesForModel(m.provider, m.model);
return {
...m,
fullModel,
routedModel,
alias: modelAliases[fullModel] || m.model,
caps: { vision: c.vision, search: c.search, reasoning: c.reasoning },
caps: {
vision: c.vision,
search: c.search,
reasoning: c.reasoning,
contextWindow: c.contextWindow,
maxOutput: c.maxOutput,
},
};
});
src/app/api/oauth/[provider]/[action]/route.js
+7 -4
View File
@@ -157,6 +157,7 @@ export async function GET(request, { params }) {
const noPkceDeviceProviders = [
"github",
"kiro",
"kimi",
"kimi-coding",
"kilocode",
"codebuddy-cn",
@@ -291,10 +292,11 @@ export async function POST(request, { params }) {
}
// Providers that don't use PKCE for device code
const noPkceProviders = ["github", "kimi-coding", "kilocode", "codebuddy-cn"];
const noPkceProviders = ["github", "kimi", "kimi-coding", "kilocode", "codebuddy-cn"];
let result;
if (noPkceProviders.includes(provider)) {
result = await pollForToken(provider, deviceCode);
// kimi needs extraData._kimiDeviceId for stable X-Msh-Device-Id (CLIProxyAPI parity)
result = await pollForToken(provider, deviceCode, null, extraData);
} else if (provider === "kiro") {
// Kiro needs extraData (clientId, clientSecret) from device code response
result = await pollForToken(provider, deviceCode, null, extraData);
@@ -315,9 +317,10 @@ export async function POST(request, { params }) {
}
if (result.success) {
// Save to database
// Save to database (legacy kimi-coding OAuth → dual-auth kimi)
const providerId = provider === "kimi-coding" ? "kimi" : provider;
const connection = await createProviderConnection({
provider,
provider: providerId,
authType: "oauth",
ownerId: user.id,
...result.tokens,
src/app/api/providers/[id]/models/route.js
+44 -7
View File
@@ -3,7 +3,7 @@ import { getProviderConnectionById } from "@/models";
import { getProviderConnectionAccess } from "@/lib/providers/connectionAccess";
import { isOpenAICompatibleProvider, isAnthropicCompatibleProvider } from "@/shared/constants/providers";
import { GEMINI_CONFIG } from "@/lib/oauth/constants/oauth";
import { refreshGoogleToken, updateProviderCredentials } from "@/sse/services/tokenRefresh";
import { refreshGoogleToken, refreshCodexToken, updateProviderCredentials } from "@/sse/services/tokenRefresh";
import { resolveOllamaLocalHost } from "open-sse/config/providers.js";
import { getModelsByProviderId } from "open-sse/config/providerModels.js";
import { resolveKiroModels } from "open-sse/services/kiroModels.js";
@@ -11,9 +11,17 @@ import { resolveKimchiModels } from "open-sse/services/kimchiModels.js";
import { resolveQoderModels } from "open-sse/services/qoderModels.js";
import { resolveGrokCliModels } from "open-sse/services/grokCliModels.js";
import { resolveConnectionProxyConfig } from "@/lib/network/connectionProxy";
import { resolveCursorModels } from "open-sse/services/cursorModels.js";
const GEMINI_CLI_MODELS_URL = "https://cloudcode-pa.googleapis.com/v1internal:fetchAvailableModels";
// The /codex/models endpoint gates each entry by minimal_client_version against this
// value, and codex CLI's own manifest (openai/codex codex-rs/models-manager/models.json)
// already requires 0.144.0 for its newest models, so a stale client_version here comes
// back 200 with those entries quietly missing instead of erroring.
const CODEX_CLIENT_VERSION = "0.144.6";
const CODEX_MODELS_URL = `https://chatgpt.com/backend-api/codex/models?client_version=${CODEX_CLIENT_VERSION}`;
const parseOpenAIStyleModels = (data) => {
if (Array.isArray(data)) return data;
return data?.data || data?.models || data?.results || [];
@@ -158,12 +166,20 @@ const PROVIDER_MODELS_CONFIG = {
parseResponse: (data) => data.data || []
},
codex: {
url: "https://chatgpt.com/backend-api/codex/models?client_version=1.0.0",
method: "GET",
headers: { "Content-Type": "application/json", "Accept": "application/json" },
authHeader: "Authorization",
authPrefix: "Bearer ",
parseResponse: parseCodexModels
customResolver: buildOAuthResolver({
refreshFn: (conn) => refreshCodexToken(conn.refreshToken),
fetchFn: (token) => fetch(CODEX_MODELS_URL, {
method: "GET",
headers: {
"Content-Type": "application/json",
"Accept": "application/json",
"Authorization": `Bearer ${token}`,
"originator": "codex_cli_rs"
}
}),
parseFn: parseCodexModels,
errorLabel: "Failed to fetch Codex models"
})
},
antigravity: {
url: "https://daily-cloudcode-pa.sandbox.googleapis.com/v1internal:models",
@@ -230,6 +246,14 @@ const PROVIDER_MODELS_CONFIG = {
authPrefix: "Bearer ",
parseResponse: (data) => data.data || []
},
"alims-intl": {
url: "https://dashscope-intl.aliyuncs.com/compatible-mode/v1/models",
method: "GET",
headers: { "Content-Type": "application/json" },
authHeader: "Authorization",
authPrefix: "Bearer ",
parseResponse: (data) => data.data || []
},
"volcengine-ark": createOpenAIModelsConfig("https://ark.cn-beijing.volces.com/api/coding/v3/models"),
byteplus: createOpenAIModelsConfig("https://ark.ap-southeast.bytepluses.com/api/coding/v3/models"),
@@ -270,6 +294,19 @@ const PROVIDER_MODELS_CONFIG = {
};
}
},
cursor: {
customResolver: async (connection) => {
const result = await resolveCursorModels({
accessToken: connection.accessToken,
providerSpecificData: connection.providerSpecificData || {},
}, { forceRefresh: true, log: console });
if (result?.models?.length) return { models: result.models };
return {
models: getStaticProviderModels("cursor"),
warning: "Cursor returned no live models; falling back to static catalog.",
};
},
},
// Custom resolvers (non-OpenAI-shaped APIs / token-refresh flows)
kiro: {
src/app/api/providers/[id]/test/testUtils.js
+7 -3
View File
@@ -76,7 +76,8 @@ const OAUTH_TEST_CONFIG = {
authPrefix: "Bearer ",
refreshable: false,
},
"kimi-coding": { checkExpiry: true, refreshable: false },
kimi: { checkExpiry: true, refreshable: true },
"kimi-coding": { checkExpiry: true, refreshable: true },
cursor: { tokenExists: true },
kilocode: {
url: `${KILOCODE_CONFIG.apiBaseUrl}/api/profile`,
@@ -627,10 +628,13 @@ async function testApiKeyConnection(connection, effectiveProxy = null) {
return { valid, error: valid ? null : "Invalid API key" };
}
case "alicode":
case "alicode-intl": {
// Aliyun Coding Plan uses OpenAI-compatible API
case "alicode-intl":
case "alims-intl": {
// Aliyun Coding Plan uses OpenAI-compatible API; alims-intl uses Model Studio compatible-mode
const aliBaseUrl = connection.provider === "alicode-intl"
? "https://coding-intl.dashscope.aliyuncs.com/v1/chat/completions"
: connection.provider === "alims-intl"
? "https://dashscope-intl.aliyuncs.com/compatible-mode/v1/chat/completions"
: "https://coding.dashscope.aliyuncs.com/v1/chat/completions";
const res = await fetchWithConnectionProxy(aliBaseUrl, {
method: "POST",
src/app/api/providers/validate/route.js
+2 -1
View File
@@ -311,11 +311,12 @@ export async function POST(request) {
case "minimax":
case "minimax-cn":
case "alicode-intl":
case "alims-intl":
case "alicode":
case "agentrouter": {
// Use baseUrl from PROVIDERS (DRY); separate openai-format vs claude-format flow
const cfg = PROVIDERS[provider];
const isOpenAiFormat = provider === "glm-cn" || provider === "alicode" || provider === "alicode-intl";
const isOpenAiFormat = provider === "glm-cn" || provider === "alicode" || provider === "alicode-intl" || provider === "alims-intl";
if (isOpenAiFormat) {
const testModel = getDefaultModel(provider);
src/app/api/v1/models/route.js
+8
View File
@@ -13,6 +13,7 @@ import { resolveQoderModels } from "open-sse/services/qoderModels.js";
import { resolveCopilotModels } from "open-sse/services/copilotModels.js";
import { resolveClinepassModels } from "open-sse/services/clinepassModels.js";
import { resolveGrokCliModels } from "open-sse/services/grokCliModels.js";
import { resolveCursorModels } from "open-sse/services/cursorModels.js";
import { updateProviderCredentials } from "@/sse/services/tokenRefresh";
import { resolveConnectionProxyConfig } from "@/lib/network/connectionProxy";
import { capabilitiesFromServiceKind, getCapabilitiesForModel } from "open-sse/providers/capabilities.js";
@@ -97,6 +98,13 @@ const LIVE_MODEL_RESOLVERS = {
});
return result?.models?.length ? { models: result.models } : null;
},
cursor: async (conn) => {
const result = await resolveCursorModels({
accessToken: conn.accessToken,
providerSpecificData: conn.providerSpecificData || {},
}, { log: console });
return result?.models?.length ? { models: result.models } : null;
}
};
const parseOpenAIStyleModels = (data) => {
src/app/globals.css
+62
View File
@@ -470,6 +470,68 @@ button:disabled,
opacity: 1;
}
/* Topology: router card pulse (no flying rings) + electric edge beam */
@keyframes topology-router-pulse {
0%, 100% {
transform: scale(1);
box-shadow:
0 0 8px #fde047,
0 0 20px color-mix(in srgb, var(--color-primary) 65%, transparent),
0 0 32px rgba(34, 211, 238, 0.4),
inset 0 0 10px rgba(253, 224, 71, 0.3);
}
50% {
transform: scale(1.06);
box-shadow:
0 0 14px #fef08a,
0 0 28px #facc15,
0 0 44px rgba(34, 211, 238, 0.6),
inset 0 0 14px rgba(255, 255, 255, 0.25);
}
}
@keyframes topology-router-icon-shake {
0%, 100% { transform: rotate(0deg) scale(1); filter: drop-shadow(0 0 2px #fde047); }
25% { transform: rotate(-5deg) scale(1.08); filter: drop-shadow(0 0 6px #facc15); }
75% { transform: rotate(5deg) scale(1.08); filter: drop-shadow(0 0 6px #22d3ee); }
}
@keyframes topology-router-label-flicker {
0%, 100% { text-shadow: 0 0 6px #fde047, 0 0 12px #22d3ee; }
50% { text-shadow: 0 0 10px #fff, 0 0 18px #facc15; }
}
@keyframes topology-edge-dash {
to { stroke-dashoffset: -36; }
}
@keyframes topology-edge-flicker {
0%, 100% { opacity: 0.85; }
40% { opacity: 1; }
55% { opacity: 0.55; }
70% { opacity: 1; }
}
.topology-router-core {
animation: topology-router-pulse 0.75s ease-in-out infinite;
border-color: #fde047 !important;
}
.topology-router-icon {
animation: topology-router-icon-shake 0.45s ease-in-out infinite;
}
.topology-router-label {
animation: topology-router-label-flicker 0.7s ease-in-out infinite;
}
.topology-router-badge {
box-shadow: 0 0 10px #fde047, 0 0 16px rgba(34, 211, 238, 0.6);
}
.topology-edge-kame {
stroke-dasharray: 8 6;
animation: topology-edge-dash 0.22s linear infinite;
}
.topology-edge-halo {
animation: topology-edge-flicker 0.18s steps(2) infinite;
}
.topology-edge-plasma {
stroke-dasharray: 14 10;
animation: topology-edge-dash 0.18s linear infinite, topology-edge-flicker 0.22s steps(3) infinite;
}
/* React Flow controls: match app theme */
.react-flow-controls-custom {
background: var(--color-surface);
src/i18n/config.js
+5
View File
@@ -20,6 +20,7 @@ export const LOCALES = [
"uk",
"tl",
"id",
"km",
"th",
"hi",
"bn",
@@ -60,6 +61,7 @@ export const LOCALE_NAMES = {
tl: "Tagalog",
id: "Indonesia",
th: "ไทย",
km: "ខ្មែរ",
hi: "हिन्दी",
bn: "বাংলা",
ur: "اردو",
@@ -141,6 +143,9 @@ export function normalizeLocale(locale) {
if (locale === "th") {
return "th";
}
if (locale === "km") {
return "km";
}
if (locale === "hi") {
return "hi";
}
src/lib/db/adapters/betterSqliteAdapter.js
+3 -3
View File
@@ -40,9 +40,9 @@ export function createBetterSqliteAdapter(filePath) {
return {
driver: "better-sqlite3",
run(sql, params = []) { return prepare(sql).run(params); },
get(sql, params = []) { return prepare(sql).get(params); },
all(sql, params = []) { return prepare(sql).all(params); },
run(sql, params = []) { return prepare(sql).run(...params); },
get(sql, params = []) { return prepare(sql).get(...params); },
all(sql, params = []) { return prepare(sql).all(...params); },
exec(sql) { return db.exec(sql); },
transaction(fn) { return db.transaction(fn)(); },
checkpoint() { try { db.pragma("wal_checkpoint(TRUNCATE)"); } catch {} },
src/lib/grokBuildConfig.js
+247
View File
@@ -0,0 +1,247 @@
export const GROK_MAIN_MODEL_SLOT = "9router";
export const GROK_BUILTIN_DEFAULT = "grok-build";
export const GROK_SUBAGENT_TYPES = ["general-purpose", "explore", "plan"];
const UNSET_SENTINEL = "__9router_unset__";
const MODELS_SECTION = "models";
const SUBAGENT_MODELS_SECTION = "subagents.models";
const escapeRegExp = (value) => value.replace(/[.*+?^${}()|[\]\\]/g, "\\$&");
const tomlString = (value) => JSON.stringify(String(value));
const sectionRegExp = (section) =>
new RegExp(
`^\\[${escapeRegExp(section)}\\][ \\t]*\\r?\\n((?:(?!\\[)[^\\r\\n]*\\r?\\n?)*)`,
"m",
);
const modelSlot = (type) => `${GROK_MAIN_MODEL_SLOT}-${type}`;
const previousDefaultRegExp = /^# 9router-prev-default = "([^"]*)"[ \t]*\r?\n?/m;
const previousSubagentRegExp = (type) =>
new RegExp(
`^# 9router-prev-subagent-${escapeRegExp(type)} = "([^"]*)"[ \\t]*\\r?\\n?`,
"m",
);
function getSectionField(toml, section, key) {
const match = toml.match(sectionRegExp(section));
if (!match) return null;
const field = match[1].match(
new RegExp(`^[ \\t]*${escapeRegExp(key)}[ \\t]*=[ \\t]*"([^"]*)"`, "m"),
);
return field ? field[1] : null;
}
function getSectionNumber(toml, section, key) {
const match = toml.match(sectionRegExp(section));
if (!match) return null;
const field = match[1].match(
new RegExp(`^[ \\t]*${escapeRegExp(key)}[ \\t]*=[ \\t]*([0-9]+(?:\\.[0-9]+)?)`, "m"),
);
if (!field) return null;
const value = Number(field[1]);
return Number.isFinite(value) ? value : null;
}
function setSectionField(toml, section, key, value) {
const match = toml.match(sectionRegExp(section));
const line = `${key} = ${tomlString(value)}`;
if (!match) {
const prefix = toml.length > 0 && !toml.endsWith("\n") ? `${toml}\n` : toml;
return `${prefix}\n[${section}]\n${line}\n`;
}
const body = match[1] || "";
const fieldRegExp = new RegExp(
`^[ \\t]*${escapeRegExp(key)}[ \\t]*=[ \\t]*"[^"]*"`,
"m",
);
const nextBody = fieldRegExp.test(body)
? body.replace(fieldRegExp, line)
: `${line}\n${body}`;
return toml.replace(match[0], `[${section}]\n${nextBody}`);
}
function deleteSectionField(toml, section, key) {
const match = toml.match(sectionRegExp(section));
if (!match) return toml;
const fieldRegExp = new RegExp(
`^[ \\t]*${escapeRegExp(key)}[ \\t]*=[^\\r\\n]*\\r?\\n?`,
"m",
);
const nextBody = (match[1] || "").replace(fieldRegExp, "");
if (!nextBody.trim()) return toml.replace(match[0], "").replace(/\n{3,}/g, "\n\n");
return toml.replace(match[0], `[${section}]\n${nextBody}`);
}
function parseModelSection(toml, slot) {
const match = toml.match(sectionRegExp(`model.${slot}`));
if (!match) return null;
const body = match[1] || "";
const contextWindow = getSectionNumber(toml, `model.${slot}`, "context_window");
return {
model: getSectionField(toml, `model.${slot}`, "model"),
base_url: getSectionField(toml, `model.${slot}`, "base_url"),
name: getSectionField(toml, `model.${slot}`, "name"),
api_key: getSectionField(toml, `model.${slot}`, "api_key"),
api_backend: getSectionField(toml, `model.${slot}`, "api_backend"),
context_window: Number.isFinite(contextWindow) && contextWindow > 0 ? contextWindow : null,
raw: body,
};
}
function buildModelSection({ slot, model, baseUrl, apiKey, contextWindow, name }) {
const lines = [
`[model.${slot}]`,
`model = ${tomlString(model)}`,
`base_url = ${tomlString(baseUrl)}`,
`name = ${tomlString(name)}`,
`description = ${tomlString("Routed via 9Router gateway")}`,
`api_backend = "chat_completions"`,
];
if (apiKey) lines.push(`api_key = ${tomlString(apiKey)}`);
if (Number.isFinite(contextWindow) && contextWindow > 0) {
lines.push(`context_window = ${Math.floor(contextWindow)}`);
}
return `${lines.join("\n")}\n`;
}
function upsertModelSection(toml, config) {
const regexp = sectionRegExp(`model.${config.slot}`);
const section = buildModelSection(config);
if (regexp.test(toml)) return toml.replace(regexp, section);
const prefix = toml.length > 0 && !toml.endsWith("\n") ? `${toml}\n` : toml;
return `${prefix}\n${section}`;
}
function removeModelSection(toml, slot) {
return toml.replace(sectionRegExp(`model.${slot}`), "").replace(/\n{3,}/g, "\n\n");
}
function insertMarker(toml, marker) {
const mainSection = sectionRegExp(`model.${GROK_MAIN_MODEL_SLOT}`);
if (mainSection.test(toml)) {
return toml.replace(mainSection, (section) => `${marker}${section}`);
}
const prefix = toml.length > 0 && !toml.endsWith("\n") ? `${toml}\n` : toml;
return `${prefix}${marker}`;
}
function rememberPreviousDefault(toml) {
if (previousDefaultRegExp.test(toml)) return toml;
const current = getSectionField(toml, MODELS_SECTION, "default");
if (!current || current === GROK_MAIN_MODEL_SLOT) return toml;
return insertMarker(toml, `# 9router-prev-default = ${tomlString(current)}\n`);
}
function restorePreviousDefault(toml) {
const previous = toml.match(previousDefaultRegExp)?.[1] || GROK_BUILTIN_DEFAULT;
let next = toml.replace(previousDefaultRegExp, "");
if (getSectionField(next, MODELS_SECTION, "default") === GROK_MAIN_MODEL_SLOT) {
next = setSectionField(next, MODELS_SECTION, "default", previous);
}
return next;
}
function rememberPreviousSubagent(toml, type) {
const regexp = previousSubagentRegExp(type);
if (regexp.test(toml)) return toml;
const current = getSectionField(toml, SUBAGENT_MODELS_SECTION, type);
const previous = current == null ? UNSET_SENTINEL : current;
return insertMarker(
toml,
`# 9router-prev-subagent-${type} = ${tomlString(previous)}\n`,
);
}
function restorePreviousSubagent(toml, type) {
const regexp = previousSubagentRegExp(type);
const previous = toml.match(regexp)?.[1] || UNSET_SENTINEL;
let next = toml.replace(regexp, "");
if (getSectionField(next, SUBAGENT_MODELS_SECTION, type) !== modelSlot(type)) {
return next;
}
if (previous === UNSET_SENTINEL) {
return deleteSectionField(next, SUBAGENT_MODELS_SECTION, type);
}
return setSectionField(next, SUBAGENT_MODELS_SECTION, type, previous);
}
export function parseGrokBuildConfig(toml) {
const subagentModels = {};
const subagentMappings = {};
for (const type of GROK_SUBAGENT_TYPES) {
const mapping = getSectionField(toml, SUBAGENT_MODELS_SECTION, type);
subagentMappings[type] = mapping;
subagentModels[type] = mapping === modelSlot(type)
? parseModelSection(toml, mapping)
: null;
}
return {
model: parseModelSection(toml, GROK_MAIN_MODEL_SLOT),
default: getSectionField(toml, MODELS_SECTION, "default"),
subagentModels,
subagentMappings,
};
}
/**
* Apply main model and optional per-type subagent overrides while preserving all unrelated TOML.
* `subagentModels === undefined` leaves existing subagent config untouched for API compatibility.
*/
export function applyGrokBuildConfig(
toml,
{ baseUrl, apiKey, model, contextWindow, subagentModels },
) {
let next = rememberPreviousDefault(toml);
next = upsertModelSection(next, {
slot: GROK_MAIN_MODEL_SLOT,
model,
baseUrl,
apiKey,
contextWindow,
name: "9Router",
});
next = setSectionField(next, MODELS_SECTION, "default", GROK_MAIN_MODEL_SLOT);
if (subagentModels && typeof subagentModels === "object") {
for (const type of GROK_SUBAGENT_TYPES) {
const selected = subagentModels[type];
const slot = modelSlot(type);
if (selected?.model) {
next = rememberPreviousSubagent(next, type);
next = upsertModelSection(next, {
slot,
model: selected.model,
baseUrl,
apiKey,
contextWindow: selected.contextWindow,
name: `9Router ${type}`,
});
next = setSectionField(next, SUBAGENT_MODELS_SECTION, type, slot);
} else {
next = restorePreviousSubagent(next, type);
next = removeModelSection(next, slot);
}
}
}
return next;
}
export function resetGrokBuildConfig(toml) {
let next = toml;
for (const type of GROK_SUBAGENT_TYPES) {
next = restorePreviousSubagent(next, type);
next = removeModelSection(next, modelSlot(type));
}
next = removeModelSection(next, GROK_MAIN_MODEL_SLOT);
next = restorePreviousDefault(next);
return next.replace(/\n{3,}/g, "\n\n");
}
export function getGrokSubagentSlot(type) {
return GROK_SUBAGENT_TYPES.includes(type) ? modelSlot(type) : null;
}
src/lib/oauth/constants/oauth.js
+13 -6
View File
@@ -89,12 +89,18 @@ export const CURSOR_CONFIG = {
},
};
// Kimi Coding OAuth Configuration (Device Code Flow)
// clientId uses env override — dynamic, not stored in registry
export const KIMI_CODING_CONFIG = {
...PROVIDER_OAUTH["kimi-coding"],
clientId: process.env.KIMI_CODING_OAUTH_CLIENT_ID || REGISTRY_PROVIDERS["kimi-coding"]?.clientId,
// Kimi Code OAuth (Device Code Flow) — merged into provider id `kimi` (dual auth)
// clientId: registry first, env override for forks
export const KIMI_CONFIG = {
...PROVIDER_OAUTH["kimi"],
clientId:
process.env.KIMI_CODING_OAUTH_CLIENT_ID ||
process.env.KIMI_OAUTH_CLIENT_ID ||
REGISTRY_PROVIDERS["kimi"]?.clientId ||
PROVIDER_OAUTH["kimi"]?.clientId,
};
// Back-compat alias for any remaining KIMI_CODING_CONFIG imports
export const KIMI_CODING_CONFIG = KIMI_CONFIG;
// KiloCode OAuth Configuration (Custom Device Auth Flow)
export const KILOCODE_CONFIG = { ...PROVIDER_OAUTH["kilocode"] };
@@ -134,7 +140,8 @@ export const PROVIDERS = {
GITHUB: "github",
KIRO: "kiro",
CURSOR: "cursor",
KIMI_CODING: "kimi-coding",
KIMI: "kimi",
KIMI_CODING: "kimi",
KILOCODE: "kilocode",
CLINE: "cline",
CLINEPASS: "clinepass",
src/lib/oauth/providers.js
+41 -14
View File
@@ -20,7 +20,7 @@ import {
KIRO_CONFIG,
assertValidAwsRegion,
CURSOR_CONFIG,
KIMI_CODING_CONFIG,
KIMI_CONFIG,
KILOCODE_CONFIG,
CLINE_CONFIG,
CLINEPASS_CONFIG,
@@ -1081,13 +1081,22 @@ const PROVIDERS = {
}),
},
"kimi-coding": {
config: KIMI_CODING_CONFIG,
// Kimi Code device flow (CLIProxyAPI internal/auth/kimi). Id is `kimi`;
// `kimi-coding` remains an alias key so old UI/API routes still resolve.
kimi: {
config: KIMI_CONFIG,
flowType: "device_code",
requestDeviceCode: async (config) => {
const { buildKimiHeaders } = await import("open-sse/config/appConstants.js");
const deviceId = crypto.randomUUID();
const headers = {
"Content-Type": "application/x-www-form-urlencoded",
Accept: "application/json",
...buildKimiHeaders(deviceId),
};
const response = await fetch(config.deviceCodeUrl, {
method: "POST",
headers: { "Content-Type": "application/x-www-form-urlencoded", Accept: "application/json" },
headers,
body: new URLSearchParams({ client_id: config.clientId }),
});
if (!response.ok) {
@@ -1095,21 +1104,30 @@ const PROVIDERS = {
throw new Error(`Device code request failed: ${error}`);
}
const data = await response.json();
const authorizeDeviceUrl = config.authorizeDeviceUrl || "https://www.kimi.com/code/authorize_device";
return {
device_code: data.device_code,
user_code: data.user_code,
verification_uri: data.verification_uri || "https://www.kimi.com/code/authorize_device",
verification_uri: data.verification_uri || authorizeDeviceUrl,
verification_uri_complete:
data.verification_uri_complete ||
`https://www.kimi.com/code/authorize_device?user_code=${data.user_code}`,
`${authorizeDeviceUrl}?user_code=${data.user_code}`,
expires_in: data.expires_in,
interval: data.interval || 5,
_kimiDeviceId: deviceId,
};
},
pollToken: async (config, deviceCode) => {
pollToken: async (config, deviceCode, _codeVerifier, extraData) => {
const { buildKimiHeaders } = await import("open-sse/config/appConstants.js");
const deviceId = extraData?._kimiDeviceId;
const headers = {
"Content-Type": "application/x-www-form-urlencoded",
Accept: "application/json",
...buildKimiHeaders(deviceId),
};
const response = await fetch(config.tokenUrl, {
method: "POST",
headers: { "Content-Type": "application/x-www-form-urlencoded", Accept: "application/json" },
headers,
body: new URLSearchParams({
grant_type: "urn:ietf:params:oauth:grant-type:device_code",
client_id: config.clientId,
@@ -1119,19 +1137,26 @@ const PROVIDERS = {
let data;
try {
data = await response.json();
} catch (e) {
const text = await response.text();
data = { error: "invalid_response", error_description: text };
} catch {
data = { error: "invalid_response", error_description: "non-json token response" };
}
return { ok: response.ok, data };
// CLIProxyAPI: Kimi returns 200 for pending states with error field
if (data.error === "authorization_pending" || data.error === "slow_down") {
return { ok: true, data };
}
if (data.access_token && deviceId) data._kimiDeviceId = deviceId;
return { ok: response.ok || !!data.access_token || !!data.error, data };
},
mapTokens: (tokens) => ({
accessToken: tokens.access_token,
refreshToken: tokens.refresh_token,
expiresIn: tokens.expires_in,
providerSpecificData: {
authMethod: "device_code",
...(tokens._kimiDeviceId ? { deviceId: tokens._kimiDeviceId } : {}),
},
}),
},
kilocode: {
config: KILOCODE_CONFIG,
flowType: "device_code",
@@ -1520,7 +1545,9 @@ const PROVIDERS = {
* Get provider handler
*/
export function getProvider(name) {
const provider = PROVIDERS[name];
// Legacy kimi-coding → kimi (dual-auth merge)
const key = name === "kimi-coding" ? "kimi" : name;
const provider = PROVIDERS[key];
if (!provider) {
throw new Error(`Unknown provider: ${name}`);
}
src/shared/components/ComboFormModal.js
+8 -6
View File
@@ -176,12 +176,14 @@ export default function ComboFormModal({ isOpen, combo, onClose, onSave, activeP
</div>
</Modal>
<ModelSelectModal isOpen={showModelSelect} onClose={() => setShowModelSelect(false)}
onSelect={handleAddModel} onDeselect={handleDeselectModel}
activeProviders={activeProviders} modelAliases={modelAliases}
availableModels={availableModels}
title="Add Model to Combo" kindFilter={kindFilter}
addedModelValues={models} closeOnSelect={false} />
{showModelSelect && (
<ModelSelectModal isOpen={showModelSelect} onClose={() => setShowModelSelect(false)}
onSelect={handleAddModel} onDeselect={handleDeselectModel}
activeProviders={activeProviders} modelAliases={modelAliases}
availableModels={availableModels}
title="Add Model to Combo" kindFilter={kindFilter}
addedModelValues={models} closeOnSelect={false} />
)}
</>
);
}
src/shared/components/Header.js
+3 -2
View File
@@ -11,6 +11,7 @@ import ThemeToggle from "@/shared/components/ThemeToggle";
import { useHeaderSearchStore } from "@/store/headerSearchStore";
import { OAUTH_PROVIDERS, APIKEY_PROVIDERS } from "@/shared/constants/config";
import { MEDIA_PROVIDER_KINDS, AI_PROVIDERS } from "@/shared/constants/providers";
import { getProviderIconSrc } from "@/shared/utils/providerIcon";
import { translate } from "@/i18n/runtime";
const getPageInfo = (pathname) => {
@@ -29,7 +30,7 @@ const getPageInfo = (pathname) => {
breadcrumbs: [
{ label: "Media Providers", href: `/dashboard/media-providers/${kindId}` },
{ label: kindConfig?.label || kindId, href: `/dashboard/media-providers/${kindId}` },
{ label: provider?.name || providerId, image: `/providers/${providerId}.png` },
{ label: provider?.name || providerId, image: getProviderIconSrc(providerId) },
],
};
}
@@ -61,7 +62,7 @@ const getPageInfo = (pathname) => {
{ label: "Providers", href: "/dashboard/providers" },
{
label: providerInfo.name,
image: `/providers/${providerInfo.id}.png`,
image: getProviderIconSrc(providerInfo.id),
},
],
};
src/shared/components/LanguageSwitcher.js
+4 -3
View File
@@ -39,6 +39,7 @@ const getLocaleInfo = (locale) => {
"tl": { name: "Tagalog", flag: "🇵🇭" },
"id": { name: "Indonesia", flag: "🇮🇩" },
"th": { name: "ไทย", flag: "🇹🇭" },
"km": { name: "ខ្មែរ", flag: "🇰🇭" },
"hi": { name: "हिन्दी", flag: "🇮🇳" },
"bn": { name: "বাংলা", flag: "🇧🇩" },
"ur": { name: "اردو", flag: "🇵🇰" },
@@ -63,9 +64,9 @@ export default function LanguageSwitcher({ className = "", isOpen: controlledOpe
const isControlled = typeof controlledOpen === "boolean";
const isOpen = isControlled ? controlledOpen : internalOpen;
const setIsOpen = (value) => {
const setIsOpen = (value, nextLocale = locale) => {
if (isControlled) {
if (!value && onClose) onClose(locale);
if (!value && onClose) onClose(nextLocale);
} else {
setInternalOpen(value);
}
@@ -92,7 +93,6 @@ export default function LanguageSwitcher({ className = "", isOpen: controlledOpe
if (nextLocale === locale || isPending) return;
setIsPending(true);
setIsOpen(false);
try {
await fetch("/api/locale", {
method: "POST",
@@ -103,6 +103,7 @@ export default function LanguageSwitcher({ className = "", isOpen: controlledOpe
// Reload translations without full page reload
await reloadTranslations();
setLocale(nextLocale);
setIsOpen(false, nextLocale);
} catch (err) {
console.error("Failed to set locale:", err);
} finally {
src/shared/components/McpMarketplaceModal.js
+1 -1
View File
@@ -154,7 +154,7 @@ export default function McpMarketplaceModal({ isOpen, onClose, onAdd, addedNames
<div className="flex items-start gap-2 px-2 py-2 hover:bg-black/5 dark:hover:bg-white/5">
{s.iconUrl ? (
// eslint-disable-next-line @next/next/no-img-element
<img src={s.iconUrl} alt="" className="size-7 rounded shrink-0 object-contain" onError={(e) => { e.target.style.display = "none"; }} />
<img src={s.iconUrl} alt="" className="size-7 rounded shrink-0 object-contain" onError={(e) => { e.target.style.display = "none"; }} loading="lazy" decoding="async" />
) : (
<div className="size-7 rounded bg-surface shrink-0" />
)}
src/shared/components/ModelSelectModal.js
+44 -2
View File
@@ -50,6 +50,46 @@ export default function ModelSelectModal({
const [providerNodes, setProviderNodes] = useState([]);
const [customModels, setCustomModels] = useState([]);
const [deletedModels, setDeletedModels] = useState({});
const [cursorModels, setCursorModels] = useState([]);
// Cursor exposes the usable catalog per account. Keep the static catalog only
// as a fallback, since it quickly becomes stale and account entitlements vary.
const cursorConnectionIds = useMemo(
() => activeProviders
.filter((provider) => provider.provider === "cursor" && provider.id)
.map((provider) => provider.id),
[activeProviders],
);
useEffect(() => {
if (!isOpen || cursorConnectionIds.length === 0) {
setCursorModels([]);
return undefined;
}
let cancelled = false;
Promise.all(cursorConnectionIds.map(async (connectionId) => {
const response = await fetch(`/api/providers/${connectionId}/models`, { cache: "no-store" });
if (!response.ok) return [];
const data = await response.json();
return Array.isArray(data.models) ? data.models : [];
}))
.then((modelLists) => {
if (cancelled) return;
const seen = new Set();
setCursorModels(modelLists.flat().filter((model) => {
if (!model?.id || seen.has(model.id)) return false;
seen.add(model.id);
return true;
}));
})
.catch((error) => {
console.warn("Unable to load Cursor models for selector:", error);
if (!cancelled) setCursorModels([]);
});
return () => { cancelled = true; };
}, [isOpen, cursorConnectionIds]);
const fetchCombos = async () => {
try {
@@ -315,7 +355,9 @@ export default function ModelSelectModal({
hasModels: mergedModels.length > 0,
};
} else {
const hardcodedModels = getModelsByProviderId(providerId);
const hardcodedModels = providerId === "cursor" && cursorModels.length > 0
? cursorModels
: getModelsByProviderId(providerId);
const hardcodedIds = new Set(hardcodedModels.map((m) => m.id));
// Custom models: if no hardcoded models (e.g. openrouter), show all aliases for this provider
@@ -387,7 +429,7 @@ export default function ModelSelectModal({
});
return groups;
}, [availableModels, filteredActiveProviders, modelAliases, allProviders, providerNodes, customModels, deletedModels, kindFilter, activeProviders]);
}, [availableModels, filteredActiveProviders, modelAliases, allProviders, providerNodes, customModels, deletedModels, kindFilter, activeProviders, cursorModels]);
// Filter combos by search query (and hide combos when kindFilter is set — combos are LLM-only by design)
const filteredCombos = useMemo(() => {
src/shared/components/OAuthModal.js
+3
View File
@@ -165,6 +165,7 @@ export default function OAuthModal({ isOpen, provider, providerInfo, onSuccess,
"github",
"qwen",
"kiro",
"kimi",
"kimi-coding",
"kilocode",
"codebuddy-cn",
@@ -210,6 +211,8 @@ export default function OAuthModal({ isOpen, provider, providerInfo, onSuccess,
_qoderMachineId: data._qoderMachineId,
_qoderVerifier: data.codeVerifier,
}
: (provider === "kimi" || provider === "kimi-coding")
? { _kimiDeviceId: data._kimiDeviceId }
: null;
startPolling(
data.device_code,
src/shared/components/ProviderIcon.js
+22 -3
View File
@@ -2,18 +2,29 @@
import { useState } from "react";
import PropTypes from "prop-types";
import { getProviderIconSrc, markProviderIconMissing } from "@/shared/utils/providerIcon";
function resolveSrc(src, providerId) {
if (providerId) return getProviderIconSrc(providerId);
if (!src) return null;
const m = String(src).match(/^\/providers\/([^/]+)\.png$/i);
if (m) return getProviderIconSrc(m[1]);
return src;
}
export default function ProviderIcon({
src,
providerId,
alt,
size = 32,
className = "",
fallbackText = "?",
fallbackColor,
}) {
const effectiveSrc = resolveSrc(src, providerId);
const [errored, setErrored] = useState(false);
if (!src || errored) {
if (!effectiveSrc || errored) {
return (
<span
className={`inline-flex items-center justify-center font-bold rounded-lg ${className}`.trim()}
@@ -31,18 +42,26 @@ export default function ProviderIcon({
return (
<img
src={src}
src={effectiveSrc}
alt={alt}
width={size}
height={size}
className={className}
onError={() => setErrored(true)}
loading="lazy"
decoding="async"
onError={() => {
const m = effectiveSrc.match(/^\/providers\/([^/]+)\.png$/i);
if (m) markProviderIconMissing(m[1]);
if (providerId) markProviderIconMissing(providerId);
setErrored(true);
}}
/>
);
}
ProviderIcon.propTypes = {
src: PropTypes.string,
providerId: PropTypes.string,
alt: PropTypes.string,
size: PropTypes.number,
className: PropTypes.string,
src/shared/constants/locales.js
+1
View File
@@ -22,6 +22,7 @@ export const LOCALE_FLAGS = {
"tl": "🇵🇭",
"id": "🇮🇩",
"th": "🇹🇭",
"km": "🇰🇭",
"hi": "🇮🇳",
"bn": "🇧🇩",
"ur": "🇵🇰",
src/shared/hooks/useModelCaps.js
+66 -30
View File
@@ -1,44 +1,80 @@
"use client";
import { useState, useEffect } from "react";
import { useState, useEffect, useCallback } from "react";
import { getCapabilitiesForModel } from "open-sse/providers/capabilities.js";
// Fetch model capabilities once and expose a lookup by fullModel ("provider/model") or bare model id.
// Module cache: one /api/models fetch shared by every useModelCaps instance.
let cache = null; // { byFull, byId } | null
let inflight = null;
function buildMaps(models) {
const byFull = {};
const byId = {};
for (const m of models || []) {
if (!m.caps) continue;
if (m.fullModel) byFull[m.fullModel] = m.caps;
if (m.routedModel) byFull[m.routedModel] = m.caps;
if (m.model) byId[m.model] = m.caps;
}
return { byFull, byId };
}
function loadModelCaps() {
if (cache) return Promise.resolve(cache);
if (inflight) return inflight;
inflight = fetch("/api/models")
.then(async (res) => {
if (!res.ok) throw new Error(`models ${res.status}`);
const data = await res.json();
cache = buildMaps(data.models);
return cache;
})
.catch(() => {
// Keep null so a later mount can retry
return { byFull: {}, byId: {} };
})
.finally(() => { inflight = null; });
return inflight;
}
// Resolve caps from a "provider/model" string or a bare model id.
function resolveCaps(byFull, byId, key) {
if (!key) return null;
if (byFull[key]) return byFull[key];
const bare = key.includes("/") ? key.slice(key.indexOf("/") + 1) : key;
if (byId[bare]) return byId[bare];
const provider = key.includes("/") ? key.slice(0, key.indexOf("/")) : null;
const c = getCapabilitiesForModel(provider, bare);
return {
vision: c.vision,
search: c.search,
reasoning: c.reasoning,
contextWindow: c.contextWindow,
maxOutput: c.maxOutput,
};
}
export function useModelCaps() {
const [byFull, setByFull] = useState({});
const [byId, setById] = useState({});
const [byFull, setByFull] = useState(() => cache?.byFull || {});
const [byId, setById] = useState(() => cache?.byId || {});
useEffect(() => {
if (cache) {
setByFull(cache.byFull);
setById(cache.byId);
return;
}
let alive = true;
(async () => {
try {
const res = await fetch("/api/models");
if (!res.ok) return;
const data = await res.json();
const full = {};
const id = {};
for (const m of data.models || []) {
if (!m.caps) continue;
if (m.fullModel) full[m.fullModel] = m.caps;
if (m.model) id[m.model] = m.caps;
}
if (alive) { setByFull(full); setById(id); }
} catch { /* ignore */ }
})();
loadModelCaps().then((maps) => {
if (alive) { setByFull(maps.byFull); setById(maps.byId); }
});
return () => { alive = false; };
}, []);
// Resolve caps from a "provider/model" string or a bare model id.
const getCaps = (key) => {
if (!key) return null;
if (byFull[key]) return byFull[key];
const bare = key.includes("/") ? key.slice(key.indexOf("/") + 1) : key;
if (byId[bare]) return byId[bare];
// Fallback: compute caps for dynamic models (passthrough/custom/suggested) not in static list
const provider = key.includes("/") ? key.slice(0, key.indexOf("/")) : null;
const c = getCapabilitiesForModel(provider, bare);
return { vision: c.vision, search: c.search, reasoning: c.reasoning };
};
const getCaps = useCallback(
(key) => resolveCaps(byFull, byId, key),
[byFull, byId],
);
return { getCaps };
}
src/shared/utils/index.js
+1
View File
@@ -1,6 +1,7 @@
// Shared Utils - Export all
export { cn } from "./cn";
export * as api from "./api";
export { getProviderIconSrc, markProviderIconMissing, resolveProviderIconId } from "./providerIcon";
import { v4 as uuidv4 } from "uuid";
src/shared/utils/providerIcon.js
+40
View File
@@ -0,0 +1,40 @@
// Provider icon paths under /public/providers.
// Alias related brands; session-cache 404s so one miss never spams again.
const ICON_ALIASES = {
"perplexity-agent": "perplexity",
"gitlab-duo": "gitlab",
"vercel-ai-gateway": "vercel",
};
// Runtime only — first 404 remembers id for the whole session
const failedIds = new Set();
function normalizeId(providerId) {
if (!providerId || typeof providerId !== "string") return "";
return providerId.trim().toLowerCase();
}
/** Resolve icon file id (after alias). Empty if previously failed this session. */
export function resolveProviderIconId(providerId) {
const id = normalizeId(providerId);
if (!id) return "";
if (failedIds.has(id)) return "";
const aliased = ICON_ALIASES[id] || id;
if (failedIds.has(aliased)) return "";
return aliased;
}
/** `/providers/{id}.png` or null when previously failed. */
export function getProviderIconSrc(providerId) {
const id = resolveProviderIconId(providerId);
return id ? `/providers/${id}.png` : null;
}
/** Call from img onError so later mounts skip the request. */
export function markProviderIconMissing(providerId) {
const id = normalizeId(providerId);
if (id) failedIds.add(id);
const aliased = ICON_ALIASES[id];
if (aliased) failedIds.add(aliased);
}
tests/__baseline__/alias-baseline.json
+7 -3
View File
@@ -10,7 +10,7 @@
"kr": "kiro",
"cu": "cursor",
"kc": "kilocode",
"kmc": "kimi-coding",
"kmc": "kimi",
"cl": "cline",
"oc": "opencode",
"ocg": "opencode-go",
@@ -98,6 +98,7 @@
"idToAlias": {
"alicode": "alicode",
"alicode-intl": "alicode-intl",
"alims-intl": "alims-intl",
"anthropic": "anthropic",
"antigravity": "ag",
"assemblyai": "assemblyai",
@@ -117,6 +118,7 @@
"cursor": "cu",
"deepgram": "deepgram",
"deepseek": "deepseek",
"featherless": "featherless",
"fireworks": "fireworks",
"gemini": "gemini",
"gemini-cli": "gc",
@@ -132,7 +134,6 @@
"kilocode": "kc",
"kimchi": "kimchi",
"kimi": "kimi",
"kimi-coding": "kmc",
"kiro": "kr",
"mimo-free": "mmf",
"minimax": "minimax",
@@ -149,6 +150,7 @@
"opencode-go": "opencode-go",
"openrouter": "openrouter",
"perplexity": "perplexity",
"perplexity-agent": "perplexity-agent",
"perplexity-web": "perplexity-web",
"qoder": "qd",
"qwen": "qw",
@@ -167,6 +169,7 @@
"ag",
"alicode",
"alicode-intl",
"alims-intl",
"anthropic",
"assemblyai",
"black-forest-labs",
@@ -188,6 +191,7 @@
"edge-tts",
"elevenlabs-tts-models",
"fal-ai",
"featherless",
"fireworks",
"gc",
"gcli",
@@ -206,7 +210,6 @@
"kc",
"kimchi",
"kimi",
"kmc",
"kr",
"local-device",
"minimax",
@@ -226,6 +229,7 @@
"openrouter-tts-models",
"openrouter-tts-voices",
"perplexity",
"perplexity-agent",
"perplexity-web",
"qd",
"qw",
tests/__baseline__/oauth-urls-baseline.json
+3 -3
View File
@@ -35,14 +35,14 @@
"xai": "https://auth.x.ai/oauth2/token",
"grok-cli": "https://auth.x.ai/oauth2/token",
"cline": "https://api.cline.bot/api/v1/auth/token",
"kimi-coding": "https://auth.kimi.com/api/oauth/token"
"kimi": "https://auth.kimi.com/api/oauth/token"
},
"authUrls": {
"kiro": "https://prod.us-east-1.auth.desktop.kiro.dev"
},
"refreshUrls": {
"cline": "https://api.cline.bot/api/v1/auth/refresh",
"kimi-coding": "https://auth.kimi.com/api/oauth/token",
"kimi": "https://auth.kimi.com/api/oauth/token",
"xai": "https://auth.x.ai/oauth2/token",
"grok-cli": "https://auth.x.ai/oauth2/token"
},
@@ -51,7 +51,7 @@
"codex": "app_EMoamEEZ73f0CkXaXp7hrann",
"qwen": "f0304373b74a44d2b584a3fb70ca9e56",
"iflow": "10009311001",
"kimi-coding": "17e5f671-d194-4dfb-9706-5516cb48c098",
"kimi": "17e5f671-d194-4dfb-9706-5516cb48c098",
"grok-cli": "b1a00492-073a-47ea-816f-4c329264a828"
}
}
tests/__baseline__/providers-baseline.json
+34 -54
View File
@@ -1,6 +1,6 @@
{
"alicode-intl": {
"baseUrl": "https://coding-intl.dashscope.aliyuncs.com/v1/chat/completions",
"baseUrl": "https://dashscope-intl.aliyuncs.com/compatible-mode/v1/chat/completions",
"headers": {},
"quirks": {
"preserveCacheControl": true
@@ -19,18 +19,17 @@
"baseUrl": "https://api.anthropic.com/v1/messages",
"format": "claude",
"headers": {
"Anthropic-Version": "2023-06-01",
"anthropic-version": "2023-06-01",
"Anthropic-Beta": "claude-code-20250219,interleaved-thinking-2025-05-14"
}
},
"antigravity": {
"baseUrls": [
"https://daily-cloudcode-pa.googleapis.com",
"https://daily-cloudcode-pa.sandbox.googleapis.com"
"https://cloudcode-pa.googleapis.com"
],
"format": "antigravity",
"headers": {
"User-Agent": "antigravity/1.107.0 darwin/arm64"
"User-Agent": "antigravity/ide/2.1.1 darwin/arm64"
},
"retry": {
"429": {
@@ -232,7 +231,7 @@
"Content-Type": "application/connect+proto",
"User-Agent": "connect-es/1.6.1"
},
"clientVersion": "3.1.0"
"clientVersion": "3.12.17"
},
"deepgram": {
"baseUrl": "https://api.deepgram.com/v1/listen",
@@ -270,6 +269,11 @@
}
]
},
"featherless": {
"baseUrl": "https://api.featherless.ai/v1/chat/completions",
"validateUrl": "https://api.featherless.ai/v1/models",
"format": "openai"
},
"fireworks": {
"baseUrl": "https://api.fireworks.ai/inference/v1/chat/completions",
"validateUrl": "https://api.fireworks.ai/inference/v1/models",
@@ -307,6 +311,7 @@
"github": {
"baseUrl": "https://api.githubcopilot.com/chat/completions",
"responsesUrl": "https://api.githubcopilot.com/responses",
"messagesUrl": "https://api.githubcopilot.com/v1/messages",
"headers": {
"copilot-integration-id": "vscode-chat",
"editor-version": "vscode/1.110.0",
@@ -398,17 +403,18 @@
"modelsUrl": "https://cli-chat-proxy.grok.com/v1/models",
"userUrl": "https://cli-chat-proxy.grok.com/v1/user",
"billingUrl": "https://cli-chat-proxy.grok.com/v1/billing",
"clientVersion": "0.2.93",
"clientIdentifier": "grok-pager",
"clientVersion": "0.2.99",
"clientIdentifier": "grok-shell",
"tokenAuth": "xai-grok-cli",
"headers": {
"User-Agent": "grok-pager/0.2.93 grok-shell/0.2.93 (linux; x86_64)",
"x-xai-token-auth": "xai-grok-cli",
"x-grok-client-identifier": "grok-pager",
"x-grok-client-version": "0.2.93",
"x-authenticateresponse": "authenticate-response"
"User-Agent": "grok-shell/0.2.99 (linux; x86_64)",
"x-grok-client-identifier": "grok-shell",
"x-grok-client-version": "0.2.99"
},
"usage": {
"url": "https://cli-chat-proxy.grok.com/v1/billing?format=credits",
"userUrl": "https://cli-chat-proxy.grok.com/v1/user?include=subscription"
},
"compactionAt": 400000,
"retry": {
"429": {
"attempts": 2,
@@ -477,7 +483,7 @@
"scheme": "bearer"
}
},
"kimi-coding": {
"kimi": {
"baseUrl": "https://api.kimi.com/coding/v1/messages",
"format": "claude",
"urlSuffix": "?beta=true",
@@ -528,45 +534,6 @@
}
]
},
"kimi": {
"baseUrl": "https://api.kimi.com/coding/v1/messages",
"format": "claude",
"urlSuffix": "?beta=true",
"headers": {
"Anthropic-Version": "2023-06-01",
"Anthropic-Beta": "claude-code-20250219,interleaved-thinking-2025-05-14"
},
"auth": {
"combined": true,
"header": "x-api-key",
"scheme": "raw"
},
"transports": [
{
"format": "openai",
"baseUrl": "https://api.kimi.com/coding/v1/chat/completions",
"auth": {
"combined": true,
"header": "Authorization",
"scheme": "bearer"
}
},
{
"format": "claude",
"baseUrl": "https://api.kimi.com/coding/v1/messages",
"urlSuffix": "?beta=true",
"headers": {
"Anthropic-Version": "2023-06-01",
"Anthropic-Beta": "claude-code-20250219,interleaved-thinking-2025-05-14"
},
"auth": {
"combined": true,
"header": "x-api-key",
"scheme": "raw"
}
}
]
},
"kiro": {
"baseUrl": "https://runtime.us-east-1.kiro.dev/generateAssistantResponse",
"baseUrls": [
@@ -774,6 +741,11 @@
"validateUrl": "https://api.perplexity.ai/models",
"format": "openai"
},
"perplexity-agent": {
"baseUrl": "https://api.perplexity.ai/v1/responses",
"validateUrl": "https://api.perplexity.ai/v1/models",
"format": "openai-responses"
},
"qoder": {
"baseUrl": "https://api3.qoder.sh/algo/api/v2/service/pro/sse/agent_chat_generation",
"headers": {},
@@ -900,5 +872,13 @@
}
}
]
},
"alims-intl": {
"baseUrl": "https://dashscope-intl.aliyuncs.com/compatible-mode/v1/chat/completions",
"headers": {},
"quirks": {
"preserveCacheControl": true
},
"format": "openai"
}
}
tests/__baseline__/verify-oauth-urls.mjs
+3 -3
View File
@@ -22,7 +22,7 @@ const resolved = {
// Grok CLI injects oauth.tokenUrl onto PROVIDERS via OAUTH_INJECT_FIELDS
"grok-cli": PROVIDERS["grok-cli"]?.tokenUrl,
cline: PROVIDERS.cline?.tokenUrl,
"kimi-coding": PROVIDERS["kimi-coding"]?.tokenUrl,
kimi: PROVIDERS.kimi?.tokenUrl,
},
authUrls: {
qwen: PROVIDERS.qwen?.authUrl,
@@ -31,7 +31,7 @@ const resolved = {
},
refreshUrls: {
cline: PROVIDERS.cline?.refreshUrl,
"kimi-coding": PROVIDERS["kimi-coding"]?.refreshUrl,
kimi: PROVIDERS.kimi?.refreshUrl,
xai: PROVIDERS.xai?.refreshUrl,
"grok-cli": PROVIDERS["grok-cli"]?.tokenUrl,
},
@@ -40,7 +40,7 @@ const resolved = {
codex: PROVIDERS.codex?.clientId,
qwen: PROVIDERS.qwen?.clientId,
iflow: PROVIDERS.iflow?.clientId,
"kimi-coding": PROVIDERS["kimi-coding"]?.clientId,
kimi: PROVIDERS.kimi?.clientId,
"grok-cli": PROVIDERS["grok-cli"]?.clientId,
},
};
tests/translator/bugs-kiro.test.js
+26
View File
@@ -5,8 +5,34 @@ import { translateRequest } from "../../open-sse/translator/index.js";
import { FORMATS } from "../../open-sse/translator/formats.js";
const O2K = (body) => translateRequest(FORMATS.OPENAI, FORMATS.KIRO, "m", body, true, null, "kiro");
const R2K = (model, body) => translateRequest(
FORMATS.OPENAI_RESPONSES,
FORMATS.KIRO,
model,
body,
true,
null,
"kiro"
);
describe("OpenAI → Kiro", () => {
it.each([
["high", "gpt-5.6-sol"],
["medium", "gpt-5.6-terra"],
["low", "gpt-5.6-luna"],
])("preserves Responses reasoning.effort %s through the full Kiro route", (effort, model) => {
const out = R2K(model, {
input: "Use the requested effort",
reasoning: { effort },
});
expect(out.additionalModelRequestFields).toEqual({
reasoning: { effort },
});
expect(out.systemPrompt || "").not.toContain("<thinking_mode>");
expect(out.systemPrompt || "").not.toContain("<max_thinking_length>");
});
// openai-to-kiro.js — safeJSONParse guards bad tool-call JSON (fixed in PR #1582)
it("malformed tool arguments do not throw the whole request", () => {
expect(() =>
tests/translator/claude-kiro-direct.test.js
+53
View File
@@ -112,6 +112,59 @@ describe("Claude → Kiro (direct route)", () => {
expect(out.systemPrompt).toContain("<max_thinking_length>24576</max_thinking_length>");
});
it("maps Claude-format effort to GPT-5.6 reasoning fields without legacy prompt tags", () => {
const out = C2K({
output_config: { effort: "low" },
messages: [{ role: "user", content: "think lightly" }],
}, null, "gpt-5.6-sol");
expect(out.additionalModelRequestFields).toEqual({
reasoning: { effort: "low" },
});
expect(out.systemPrompt || "").not.toContain("<thinking_mode>");
expect(out.systemPrompt || "").not.toContain("<max_thinking_length>");
});
it.each(["auto", "minimal", "ultra"])(
"keeps the legacy thinking fallback for unsupported GPT-5.6 effort %s",
(effort) => {
const out = C2K({
output_config: { effort },
messages: [{ role: "user", content: "Use legacy thinking" }],
}, null, "gpt-5.6-sol");
expect(out.additionalModelRequestFields).toBeUndefined();
expect(out.systemPrompt).toContain("<thinking_mode>enabled</thinking_mode>");
expect(out.systemPrompt).toContain("<max_thinking_length>");
}
);
it.each(["none", "off", "disabled"])(
"keeps GPT-5.6 reasoning intentionally disabled for effort %s",
(effort) => {
const out = C2K({
output_config: { effort },
messages: [{ role: "user", content: "Do not reason" }],
}, null, "gpt-5.6-sol");
expect(out.additionalModelRequestFields).toBeUndefined();
expect(out.systemPrompt || "").not.toContain("<thinking_mode>");
expect(out.systemPrompt || "").not.toContain("<max_thinking_length>");
}
);
it("keeps explicit Claude effort ahead of an injected OpenAI effort", () => {
const out = C2K({
output_config: { effort: "low" },
reasoning_effort: "high",
messages: [{ role: "user", content: "honor the client effort" }],
}, null, "gpt-5.6-sol");
expect(out.additionalModelRequestFields).toEqual({
reasoning: { effort: "low" },
});
});
it("sends Claude system as top-level systemPrompt and keeps a user-content fallback", () => {
const out = C2K({
system: "system-only instruction",
tests/unit/alicode-intl-endpoint-2591.test.js
+22 -11
View File
@@ -1,24 +1,35 @@
// #2591 — Alibaba Intl (alicode-intl) must use the OpenAI-compatible-mode
// DashScope endpoint so standard DashScope API keys work. The previous
// coding-intl host only accepted Alibaba Coding Plan keys and rejected
// ordinary DashScope keys with "Invalid API key".
// #2591 — Alibaba Intl key types split across two hosts:
// - alicode-intl: Coding Plan keys (sk-sp-...) → coding-intl.dashscope.aliyuncs.com
// - alims-intl: standard DashScope API keys (sk-...) → dashscope-intl.aliyuncs.com/compatible-mode
// The two key types are NOT interchangeable across hosts. Split into two providers
// so each key type reaches its own host.
import { describe, it, expect } from "vitest";
import alicodeIntl from "../../open-sse/providers/registry/alicode-intl.js";
import alimsIntl from "../../open-sse/providers/registry/alims-intl.js";
describe("alicode-intl endpoint (issue #2591)", () => {
it("routes to the compatible-mode DashScope endpoint", () => {
describe("alicode-intl endpoint (Coding Plan keys)", () => {
it("routes to the coding-intl host for Coding Plan keys", () => {
expect(alicodeIntl.id).toBe("alicode-intl");
expect(alicodeIntl.transport.baseUrl).toBe(
"https://dashscope-intl.aliyuncs.com/compatible-mode/v1/chat/completions"
"https://coding-intl.dashscope.aliyuncs.com/v1/chat/completions"
);
});
it("does not use the coding-intl host that rejects standard keys", () => {
expect(alicodeIntl.transport.baseUrl).not.toContain("coding-intl.dashscope.aliyuncs.com");
});
it("keeps the chat/completions path and preserveCacheControl quirk", () => {
expect(alicodeIntl.transport.baseUrl).toContain("/v1/chat/completions");
expect(alicodeIntl.transport.quirks.preserveCacheControl).toBe(true);
});
});
describe("alims-intl endpoint (standard DashScope keys)", () => {
it("routes to the compatible-mode DashScope endpoint for standard keys", () => {
expect(alimsIntl.id).toBe("alims-intl");
expect(alimsIntl.transport.baseUrl).toBe(
"https://dashscope-intl.aliyuncs.com/compatible-mode/v1/chat/completions"
);
});
it("does not use the coding-intl host that rejects standard keys", () => {
expect(alimsIntl.transport.baseUrl).not.toContain("coding-intl.dashscope.aliyuncs.com");
});
});
tests/unit/codex-refresh-token.test.js
+2 -1
View File
@@ -152,7 +152,8 @@ describe("Codex Refresh Token", () => {
expect(getRefreshLeadMs("claude")).toBe(4 * 60 * 60 * 1000); // 4 hours
expect(getRefreshLeadMs("iflow")).toBe(24 * 60 * 60 * 1000); // 24 hours
expect(getRefreshLeadMs("qwen")).toBe(20 * 60 * 1000); // 20 minutes
expect(getRefreshLeadMs("kimi-coding")).toBe(5 * 60 * 1000); // 5 minutes
expect(getRefreshLeadMs("kimi")).toBe(5 * 60 * 1000); // 5 minutes
expect(getRefreshLeadMs("kimi-coding")).toBe(5 * 60 * 1000); // legacy alias
expect(getRefreshLeadMs("antigravity")).toBe(5 * 60 * 1000); // 5 minutes
});
tests/unit/cursor-agent-proto.test.js
+282
View File
@@ -0,0 +1,282 @@
import { describe, expect, it } from "vitest";
import {
decodeMessage,
encodeField,
encodeAgentValue,
decodeAgentValue,
encodeMcpToolDefinition,
encodeMcpTools,
decodeMcpArgs,
encodeMcpResultSuccess,
encodeMcpResultError,
encodeMcpResultToolNotFound,
} from "../../open-sse/utils/cursorProtobuf.js";
import {
isAgentCapableRequest,
buildAgentRunFrame,
} from "../../open-sse/executors/cursor.js";
// AgentService (agent.v1) codec tests — validate the production implementation
// in cursorProtobuf.js + the executor's frame builders. Pure round-trip, no network.
// Field numbers verified against Cursor's agent.proto (extracted via @oh-my-pi).
const LEN = 2;
// McpArgs.args map entry { field1: key, field2: Value }
const entry = (k, v) => Buffer.concat([
Buffer.from(encodeField(2, LEN,
Buffer.concat([Buffer.from(encodeField(1, LEN, k)), Buffer.from(encodeField(2, LEN, encodeAgentValue(v)))])
)),
]);
describe("Cursor AgentService codec (cursorProtobuf.js)", () => {
describe("google.protobuf.Value round-trip", () => {
const cases = [
["null", null],
["bool true", true],
["bool false", false],
["string", "hello"],
["integer", 42],
["float", 3.14],
["empty object", {}],
["flat object", { a: 1, b: "x", c: true }],
["nested object", { outer: { inner: [1, 2, "three"] } }],
["array of mixed", [1, "two", false, null]],
["deeply nested", { a: { b: { c: { d: 1 } } } }],
];
for (const [label, value] of cases) {
it(`encodes/decodes ${label}`, () => {
expect(decodeAgentValue(encodeAgentValue(value))).toEqual(value);
});
}
});
describe("McpToolDefinition", () => {
it("encodes name, description, input_schema (Value), provider, tool_name", () => {
const schema = { type: "object", properties: { city: { type: "string" } }, required: ["city"] };
const def = encodeMcpToolDefinition({ function: { name: "get_weather", description: "Get weather", parameters: schema } });
const msg = decodeMessage(def);
expect(Buffer.from(msg.get(1)[0].value).toString("utf8")).toBe("get_weather");
expect(Buffer.from(msg.get(2)[0].value).toString("utf8")).toBe("Get weather");
expect(Buffer.from(msg.get(4)[0].value).toString("utf8")).toBe("9router");
expect(Buffer.from(msg.get(5)[0].value).toString("utf8")).toBe("get_weather");
expect(decodeAgentValue(msg.get(3)[0].value)).toEqual(schema);
});
it("preserves nested JSON-schema types", () => {
const schema = {
type: "object",
properties: {
query: { type: "string", description: "search query" },
opts: { type: "array", items: { type: "string" } },
},
required: ["query"],
};
const def = encodeMcpToolDefinition({ function: { name: "search", parameters: schema } });
const msg = decodeMessage(def);
expect(decodeAgentValue(msg.get(3)[0].value)).toEqual(schema);
});
it("accepts flat tool shape (no .function wrapper)", () => {
const def = encodeMcpToolDefinition({ name: "noop", description: "d", inputSchema: { type: "object" } });
const msg = decodeMessage(def);
expect(Buffer.from(msg.get(1)[0].value).toString("utf8")).toBe("noop");
});
});
describe("encodeMcpTools", () => {
it("produces empty bytes for no tools", () => {
expect(encodeMcpTools([]).length).toBe(0);
expect(encodeMcpTools().length).toBe(0);
});
it("wraps multiple tool defs as repeated field 1", () => {
const tools = [
{ function: { name: "get_weather", parameters: { type: "object" } } },
{ function: { name: "calculate", parameters: { type: "object" } } },
];
const mcpTools = encodeMcpTools(tools);
const inner = decodeMessage(mcpTools);
expect(inner.get(1).length).toBe(2);
});
});
describe("McpArgs decode", () => {
it("decodes name, toolName, toolCallId, and typed args map", () => {
const argsBytes = Buffer.concat([
entry("city", "Hanoi"),
entry("count", 5),
entry("flag", true),
entry("nested", { a: [1, 2] }),
]);
const mcpArgs = Buffer.concat([
Buffer.from(encodeField(1, LEN, "get_weather")),
argsBytes,
Buffer.from(encodeField(3, LEN, "call_abc")),
Buffer.from(encodeField(5, LEN, "get_weather")),
]);
const decoded = decodeMcpArgs(mcpArgs);
expect(decoded.name).toBe("get_weather");
expect(decoded.toolName).toBe("get_weather");
expect(decoded.toolCallId).toBe("call_abc");
expect(decoded.args).toEqual({ city: "Hanoi", count: 5, flag: true, nested: { a: [1, 2] } });
});
it("handles empty args map", () => {
const mcpArgs = Buffer.concat([
Buffer.from(encodeField(1, LEN, "noop")),
Buffer.from(encodeField(5, LEN, "noop")),
]);
expect(decodeMcpArgs(mcpArgs).args).toEqual({});
});
});
describe("McpResult success", () => {
it("builds success with single text content", () => {
const bytes = encodeMcpResultSuccess({ textItems: ['{"temp":32}'], isError: false });
const msg = decodeMessage(bytes); // McpResult level
expect(msg.has(1)).toBe(true); // success variant
const success = decodeMessage(msg.get(1)[0].value);
expect(success.get(1).length).toBe(1);
expect(success.get(2)[0].value).toBe(0); // is_error=false
const item = decodeMessage(success.get(1)[0].value);
const textContent = decodeMessage(item.get(1)[0].value);
expect(Buffer.from(textContent.get(1)[0].value).toString("utf8")).toBe('{"temp":32}');
});
it("builds success with multiple text items", () => {
const bytes = encodeMcpResultSuccess({ textItems: ["line1", "line2"] });
const success = decodeMessage(decodeMessage(bytes).get(1)[0].value);
expect(success.get(1).length).toBe(2);
});
it("marks is_error=true", () => {
const bytes = encodeMcpResultSuccess({ textItems: ["fail"], isError: true });
const success = decodeMessage(decodeMessage(bytes).get(1)[0].value);
expect(success.get(2)[0].value).toBe(1);
});
});
describe("McpResult image content", () => {
it("builds image item with raw bytes + mime type", () => {
const imgBytes = new Uint8Array([0x89, 0x50, 0x4e, 0x47]);
const bytes = encodeMcpResultSuccess({ imageItems: [{ data: imgBytes, mimeType: "image/png" }] });
const success = decodeMessage(decodeMessage(bytes).get(1)[0].value);
const item = decodeMessage(success.get(1)[0].value);
expect(item.has(2)).toBe(true); // image variant
const img = decodeMessage(item.get(2)[0].value);
expect(Buffer.from(img.get(1)[0].value)).toEqual(Buffer.from(imgBytes));
expect(Buffer.from(img.get(2)[0].value).toString("utf8")).toBe("image/png");
});
it("builds mixed text + image content", () => {
const imgBytes = new Uint8Array([1, 2, 3]);
const bytes = encodeMcpResultSuccess({ textItems: ["see image"], imageItems: [{ data: imgBytes, mimeType: "image/jpeg" }] });
const success = decodeMessage(decodeMessage(bytes).get(1)[0].value);
expect(success.get(1).length).toBe(2);
expect(decodeMessage(success.get(1)[0].value).has(1)).toBe(true); // text
expect(decodeMessage(success.get(1)[1].value).has(2)).toBe(true); // image
});
});
describe("McpResult error / toolNotFound", () => {
it("builds error result (field 2)", () => {
const bytes = encodeMcpResultError("tool crashed");
const msg = decodeMessage(bytes);
expect(msg.has(2)).toBe(true);
const err = decodeMessage(msg.get(2)[0].value);
expect(Buffer.from(err.get(1)[0].value).toString("utf8")).toBe("tool crashed");
});
it("builds toolNotFound result (field 5)", () => {
const bytes = encodeMcpResultToolNotFound("missing_tool");
const msg = decodeMessage(bytes);
expect(msg.has(5)).toBe(true);
const tnf = decodeMessage(msg.get(5)[0].value);
expect(Buffer.from(tnf.get(1)[0].value).toString("utf8")).toBe("missing_tool");
});
});
});
describe("Cursor AgentService executor helpers (cursor.js)", () => {
describe("isAgentCapableRequest", () => {
it("accepts plain text content", () => {
expect(isAgentCapableRequest({ messages: [{ role: "user", content: "hi" }] })).toBe(true);
});
it("accepts array text content", () => {
expect(isAgentCapableRequest({ messages: [{ role: "user", content: [{ type: "text", text: "hi" }] }] })).toBe(true);
});
it("accepts request with tools declared", () => {
expect(isAgentCapableRequest({ messages: [{ role: "user", content: "hi" }], tools: [{ function: { name: "t" } }] })).toBe(true);
});
it("accepts history with assistant tool_calls + tool results", () => {
expect(isAgentCapableRequest({
messages: [
{ role: "user", content: "weather?" },
{ role: "assistant", content: null, tool_calls: [{ id: "c1", type: "function", function: { name: "get_weather", arguments: "{}" } }] },
{ role: "tool", tool_call_id: "c1", content: "sunny" },
{ role: "user", content: "thanks" },
],
})).toBe(true);
});
it("rejects non-text (image) content", () => {
expect(isAgentCapableRequest({ messages: [{ role: "user", content: [{ type: "image_url" }] }] })).toBe(false);
});
it("rejects missing messages", () => {
expect(isAgentCapableRequest({})).toBe(false);
expect(isAgentCapableRequest(null)).toBe(false);
});
});
describe("buildAgentRunFrame", () => {
// buildAgentRunFrame returns a wrapped Connect-RPC frame (5-byte header + AgentClientMessage).
const unwrap = (frame) => frame.subarray(5);
it("encodes a text-only run request with system + model", () => {
const frame = unwrap(buildAgentRunFrame(
[{ role: "system", content: "be brief" }, { role: "user", content: "hi" }],
"gpt-5.2",
));
const clientMsg = decodeMessage(frame);
expect(clientMsg.has(1)).toBe(true); // run_request
const run = decodeMessage(clientMsg.get(1)[0].value);
expect(run.has(2)).toBe(true); // action
expect(run.has(9)).toBe(true); // requested_model
});
it("encodes mcp_tools (field 4) when tools are provided", () => {
const tools = [{ function: { name: "get_weather", description: "weather", parameters: { type: "object", properties: { city: { type: "string" } } } } }];
const frame = unwrap(buildAgentRunFrame([{ role: "user", content: "weather?" }], "gpt-5.2", tools));
const run = decodeMessage(decodeMessage(frame).get(1)[0].value);
expect(run.has(4)).toBe(true); // mcp_tools
const mcpTools = decodeMessage(run.get(4)[0].value);
expect(mcpTools.get(1).length).toBe(1);
});
it("omits mcp_tools when no tools provided", () => {
const frame = unwrap(buildAgentRunFrame([{ role: "user", content: "hi" }], "gpt-5.2", []));
const run = decodeMessage(decodeMessage(frame).get(1)[0].value);
expect(run.has(4)).toBe(false);
});
it("encodes conversation_history from prior turns including tool calls/results", () => {
const messages = [
{ role: "user", content: "weather in Tokyo?" },
{ role: "assistant", content: null, tool_calls: [{ id: "c1", type: "function", function: { name: "get_weather", arguments: '{"city":"Tokyo"}' } }] },
{ role: "tool", tool_call_id: "c1", content: "18C cloudy" },
{ role: "user", content: "thanks" },
];
const frame = unwrap(buildAgentRunFrame(messages, "gpt-5.2", []));
const run = decodeMessage(decodeMessage(frame).get(1)[0].value);
const action = decodeMessage(run.get(2)[0].value);
const userAction = decodeMessage(action.get(1)[0].value);
expect(userAction.has(7)).toBe(true); // conversation_history (field 7)
const history = decodeMessage(userAction.get(7)[0].value);
expect(history.get(1).length).toBeGreaterThanOrEqual(2); // prior turns
});
});
});
tests/unit/cursor-models.test.js
+103
View File
@@ -0,0 +1,103 @@
import { afterEach, beforeEach, describe, expect, it, vi } from "vitest";
import {
clearCursorModelCache,
parseCursorUsableModels,
resolveCursorModels,
} from "../../open-sse/services/cursorModels.js";
const originalFetch = global.fetch;
function varint(value) {
const bytes = [];
while (value >= 0x80) {
bytes.push((value & 0x7f) | 0x80);
value >>>= 7;
}
bytes.push(value);
return Uint8Array.from(bytes);
}
function field(fieldNumber, value) {
return Uint8Array.from([(fieldNumber << 3) | 2, ...varint(value.length), ...value]);
}
function text(value) {
return new TextEncoder().encode(value);
}
function concat(...parts) {
const size = parts.reduce((sum, part) => sum + part.length, 0);
const result = new Uint8Array(size);
let offset = 0;
for (const part of parts) {
result.set(part, offset);
offset += part.length;
}
return result;
}
function model(id, name) {
return field(1, concat(field(1, text(id)), field(4, text(name))));
}
describe("Cursor live model catalog", () => {
beforeEach(() => {
clearCursorModelCache();
});
afterEach(() => {
global.fetch = originalFetch;
clearCursorModelCache();
});
it("decodes the GetUsableModels protobuf response", () => {
const payload = concat(
model("default", "Auto"),
model("gpt-5.3-codex", "GPT 5.3 Codex"),
model("gpt-5.3-codex", "Duplicate"),
);
expect(parseCursorUsableModels(payload)).toEqual([
{ id: "default", name: "Auto" },
{ id: "gpt-5.3-codex", name: "GPT 5.3 Codex" },
]);
});
it("fetches the account-specific catalog and caches it", async () => {
const payload = concat(model("claude-4.6-opus", "Claude 4.6 Opus"));
global.fetch = vi.fn().mockResolvedValue(new Response(payload, { status: 200 }));
const credentials = {
accessToken: "cursor-token",
providerSpecificData: { machineId: "machine-id" },
};
await expect(resolveCursorModels(credentials)).resolves.toEqual({
models: [{ id: "claude-4.6-opus", name: "Claude 4.6 Opus" }],
});
await expect(resolveCursorModels(credentials)).resolves.toEqual({
models: [{ id: "claude-4.6-opus", name: "Claude 4.6 Opus" }],
});
expect(global.fetch).toHaveBeenCalledTimes(1);
expect(global.fetch).toHaveBeenCalledWith(
"https://agent.api5.cursor.sh/agent.v1.AgentService/GetUsableModels",
expect.objectContaining({
method: "POST",
body: expect.any(Uint8Array),
headers: expect.objectContaining({
"content-type": "application/proto",
accept: "application/proto",
}),
}),
);
});
it("fails open when the Cursor catalog request fails", async () => {
global.fetch = vi.fn().mockResolvedValue(new Response("no", { status: 403 }));
await expect(resolveCursorModels({
accessToken: "cursor-token",
providerSpecificData: { machineId: "machine-id" },
})).resolves.toBeNull();
});
});
tests/unit/grok-build-config.test.js
+176
View File
@@ -0,0 +1,176 @@
import { describe, expect, it } from "vitest";
import {
applyGrokBuildConfig,
getGrokSubagentSlot,
parseGrokBuildConfig,
resetGrokBuildConfig,
} from "../../src/lib/grokBuildConfig.js";
const BASE_CONFIG = `[cli]
installer = "internal"
[ui]
yolo = false
[models]
default = "grok-4.5"
default_reasoning_effort = "high"
[subagents]
enabled = true
[subagents.models]
general-purpose = "grok-4.5"
explore = "grok-build"
plan = "grok-4.5"
[mcp_servers.example]
url = "https://example.com/mcp"
enabled = true
`;
const APPLY_INPUT = {
baseUrl: "http://127.0.0.1:20128/v1",
apiKey: "sk-test",
model: "cx/gpt-5.6-sol",
contextWindow: 400000,
subagentModels: {
"general-purpose": { model: "cc/claude-sonnet-5", contextWindow: 1000000 },
explore: { model: "gemini/gemini-3-flash", contextWindow: 1048576 },
},
};
describe("grokBuildConfig", () => {
it("creates independent main and per-type subagent model slots", () => {
const result = applyGrokBuildConfig(BASE_CONFIG, APPLY_INPUT);
const parsed = parseGrokBuildConfig(result);
expect(parsed.default).toBe("9router");
expect(parsed.model).toMatchObject({
model: "cx/gpt-5.6-sol",
base_url: "http://127.0.0.1:20128/v1",
context_window: 400000,
});
expect(parsed.subagentMappings).toMatchObject({
"general-purpose": "9router-general-purpose",
explore: "9router-explore",
plan: "grok-4.5",
});
expect(parsed.subagentModels["general-purpose"]).toMatchObject({
model: "cc/claude-sonnet-5",
context_window: 1000000,
});
expect(parsed.subagentModels.explore).toMatchObject({
model: "gemini/gemini-3-flash",
context_window: 1048576,
});
expect(parsed.subagentModels.plan).toBeNull();
});
it("preserves unrelated config sections", () => {
const result = applyGrokBuildConfig(BASE_CONFIG, APPLY_INPUT);
expect(result).toContain("[cli]\ninstaller = \"internal\"");
expect(result).toContain("[ui]\nyolo = false");
expect(result).toContain("default_reasoning_effort = \"high\"");
expect(result).toContain("[mcp_servers.example]");
expect(result).toContain("url = \"https://example.com/mcp\"");
});
it("is idempotent and updates owned slots without duplicate sections", () => {
let result = applyGrokBuildConfig(BASE_CONFIG, APPLY_INPUT);
result = applyGrokBuildConfig(result, {
...APPLY_INPUT,
model: "cc/claude-opus-4.8",
contextWindow: 1000000,
subagentModels: {
...APPLY_INPUT.subagentModels,
explore: { model: "mimo/mimo", contextWindow: 262144 },
},
});
expect(result.match(/^\[model\.9router\]$/gm)).toHaveLength(1);
expect(result.match(/^\[model\.9router-general-purpose\]$/gm)).toHaveLength(1);
expect(result.match(/^\[model\.9router-explore\]$/gm)).toHaveLength(1);
expect(result.match(/^# 9router-prev-subagent-explore/gm)).toHaveLength(1);
expect(parseGrokBuildConfig(result).model).toMatchObject({
model: "cc/claude-opus-4.8",
context_window: 1000000,
});
expect(parseGrokBuildConfig(result).subagentModels.explore).toMatchObject({
model: "mimo/mimo",
context_window: 262144,
});
});
it("blank override restores previous subagent mapping and removes owned slot", () => {
let result = applyGrokBuildConfig(BASE_CONFIG, APPLY_INPUT);
result = applyGrokBuildConfig(result, {
...APPLY_INPUT,
subagentModels: {
"general-purpose": APPLY_INPUT.subagentModels["general-purpose"],
// explore omitted => inherit / restore previous
},
});
const parsed = parseGrokBuildConfig(result);
expect(parsed.subagentMappings.explore).toBe("grok-build");
expect(parsed.subagentModels.explore).toBeNull();
expect(result).not.toContain("[model.9router-explore]");
expect(parsed.subagentMappings["general-purpose"]).toBe("9router-general-purpose");
});
it("reset restores previous default and all previous subagent mappings", () => {
const applied = applyGrokBuildConfig(BASE_CONFIG, APPLY_INPUT);
const reset = resetGrokBuildConfig(applied);
const parsed = parseGrokBuildConfig(reset);
expect(parsed.default).toBe("grok-4.5");
expect(parsed.model).toBeNull();
expect(parsed.subagentMappings).toEqual({
"general-purpose": "grok-4.5",
explore: "grok-build",
plan: "grok-4.5",
});
expect(reset).not.toContain("[model.9router-");
expect(reset).not.toContain("9router-prev-");
expect(reset).toContain("[mcp_servers.example]");
});
it("removes mappings that were originally unset", () => {
const config = `[models]\ndefault = "grok-build"\n\n[mcp_servers.x]\nenabled = true\n`;
const applied = applyGrokBuildConfig(config, {
...APPLY_INPUT,
subagentModels: {
plan: { model: "cc/claude-sonnet-5", contextWindow: 1000000 },
},
});
const reset = resetGrokBuildConfig(applied);
expect(parseGrokBuildConfig(applied).subagentMappings.plan).toBe("9router-plan");
expect(parseGrokBuildConfig(reset).subagentMappings.plan).toBeNull();
expect(reset).not.toContain("[subagents.models]");
expect(reset).toContain("[mcp_servers.x]");
});
it("legacy callers without subagentModels leave existing overrides untouched", () => {
const applied = applyGrokBuildConfig(BASE_CONFIG, APPLY_INPUT);
const updatedMainOnly = applyGrokBuildConfig(applied, {
baseUrl: APPLY_INPUT.baseUrl,
apiKey: APPLY_INPUT.apiKey,
model: "gemini/gemini-3.1-pro",
contextWindow: 1048576,
});
const parsed = parseGrokBuildConfig(updatedMainOnly);
expect(parsed.model.model).toBe("gemini/gemini-3.1-pro");
expect(parsed.subagentMappings.explore).toBe("9router-explore");
expect(parsed.subagentModels.explore.model).toBe("gemini/gemini-3-flash");
});
it("returns stable slot names only for supported subagent types", () => {
expect(getGrokSubagentSlot("general-purpose")).toBe("9router-general-purpose");
expect(getGrokSubagentSlot("explore")).toBe("9router-explore");
expect(getGrokSubagentSlot("plan")).toBe("9router-plan");
expect(getGrokSubagentSlot("unknown")).toBeNull();
});
});
tests/unit/kiro-nonstream-error.test.js
+64
View File
@@ -0,0 +1,64 @@
import { describe, expect, it, vi } from "vitest";
vi.mock("@/lib/usageDb.js", () => ({
appendRequestLog: vi.fn(async () => {}),
saveRequestDetail: vi.fn(async () => {}),
saveRequestUsage: vi.fn(async () => {})
}));
const { FORMATS } = await import("../../open-sse/translator/formats.js");
const {
handleForcedSSEToJson,
parseSSEToOpenAIResponse
} = await import("../../open-sse/handlers/chatCore/sseToJsonHandler.js");
describe("Kiro non-streaming error propagation", () => {
it("prefers a terminal SSE error over earlier semantic chunks", () => {
const raw = [
'data: {"choices":[{"delta":{"content":"partial"},"finish_reason":null}]}',
'data: {"error":{"message":"Kiro transport failed","code":"kiro_missing_terminal"}}',
"data: [DONE]"
].join("\n\n");
expect(parseSSEToOpenAIResponse(raw, "kiro")).toEqual({
error: {
message: "Kiro transport failed",
code: "kiro_missing_terminal"
}
});
});
it("returns 502 instead of collapsing a failed Kiro SSE stream into stop", async () => {
const encoder = new TextEncoder();
const raw = [
'data: {"choices":[{"delta":{"content":"partial"},"finish_reason":null}]}',
'data: {"error":{"message":"Kiro stream ended incompletely","code":"kiro_missing_terminal"}}',
"data: [DONE]",
""
].join("\n\n");
const result = await handleForcedSSEToJson({
providerResponse: new Response(new ReadableStream({
start(controller) {
controller.enqueue(encoder.encode(raw));
controller.close();
}
}), { headers: { "content-type": "text/event-stream" } }),
sourceFormat: FORMATS.OPENAI,
provider: "kiro",
model: "kr/claude-opus-4.8",
body: { model: "kr/claude-opus-4.8", messages: [] },
stream: false,
requestStartTime: Date.now(),
connectionId: "test-connection",
clientRawRequest: { endpoint: "/v1/chat/completions" },
trackDone: vi.fn(),
appendLog: vi.fn()
});
const json = await result.response.json();
expect(result.success).toBe(false);
expect(result.response.status).toBe(502);
expect(json.error.message).toContain("Kiro stream ended incompletely");
expect(json).not.toHaveProperty("choices");
});
});
tests/unit/kiro-terminal-integrity.test.js
+778
View File
@@ -0,0 +1,778 @@
import { afterEach, beforeEach, describe, expect, it, vi } from "vitest";
const fetchMock = vi.fn();
vi.mock("../../open-sse/utils/proxyFetch.js", () => ({
proxyAwareFetch: (...args) => fetchMock(...args)
}));
const { KiroExecutor } = await import("../../open-sse/executors/kiro.js");
const encoder = new TextEncoder();
const credentials = {
accessToken: "test-token",
providerSpecificData: { kiroToolCallRepair: true }
};
function crc32(bytes) {
let crc = 0xffffffff;
for (const byte of bytes) {
crc ^= byte;
for (let bit = 0; bit < 8; bit++) {
crc = (crc >>> 1) ^ ((crc & 1) ? 0xedb88320 : 0);
}
}
return (crc ^ 0xffffffff) >>> 0;
}
function encodeHeader(name, value) {
const nameBytes = encoder.encode(name);
const valueBytes = encoder.encode(value);
const bytes = new Uint8Array(1 + nameBytes.length + 3 + valueBytes.length);
let offset = 0;
bytes[offset++] = nameBytes.length;
bytes.set(nameBytes, offset);
offset += nameBytes.length;
bytes[offset++] = 7;
new DataView(bytes.buffer).setUint16(offset, valueBytes.length, false);
offset += 2;
bytes.set(valueBytes, offset);
return bytes;
}
function concat(chunks) {
const output = new Uint8Array(chunks.reduce((size, chunk) => size + chunk.byteLength, 0));
let offset = 0;
for (const chunk of chunks) {
output.set(chunk, offset);
offset += chunk.byteLength;
}
return output;
}
function frameFromEntries(entries, payload) {
const headers = concat(entries.map(([name, value]) => encodeHeader(name, value)));
const payloadBytes = encoder.encode(JSON.stringify(payload));
const totalLength = 12 + headers.byteLength + payloadBytes.byteLength + 4;
const frame = new Uint8Array(totalLength);
const view = new DataView(frame.buffer);
view.setUint32(0, totalLength, false);
view.setUint32(4, headers.byteLength, false);
frame.set(headers, 12);
frame.set(payloadBytes, 12 + headers.byteLength);
return checksum(frame);
}
function frame(eventType, payload) {
return frameFromEntries([[":event-type", eventType]], payload);
}
function checksum(bytes) {
const view = new DataView(bytes.buffer, bytes.byteOffset, bytes.byteLength);
view.setUint32(8, crc32(bytes.subarray(0, 8)), false);
view.setUint32(bytes.byteLength - 4, crc32(bytes.subarray(0, bytes.byteLength - 4)), false);
return bytes;
}
function response(frames, status = 200) {
return new Response(new ReadableStream({
start(controller) {
for (const value of frames) controller.enqueue(value);
controller.close();
}
}), { status, statusText: status === 200 ? "OK" : "Upstream Error" });
}
function controlledResponse(frames = []) {
let controller;
const value = new Response(new ReadableStream({
start(streamController) {
controller = streamController;
for (const item of frames) controller.enqueue(item);
}
}), { status: 200 });
return {
value,
enqueue(item) {
controller.enqueue(item);
},
close() {
controller.close();
}
};
}
async function text(stream) {
const reader = stream.getReader();
const decoder = new TextDecoder();
let output = "";
while (true) {
const { done, value } = await reader.read();
if (done) return output + decoder.decode();
output += decoder.decode(value, { stream: true });
}
}
async function execute(executor = new KiroExecutor(), overrides = {}) {
return executor.execute({
model: "kr/claude-opus-4.8",
body: { systemPrompt: "base", conversationState: {} },
stream: true,
credentials,
...overrides
});
}
beforeEach(() => {
fetchMock.mockReset();
delete process.env.KIRO_TOOL_CALL_REPAIR_BUFFER_MAX_BYTES;
delete process.env.KIRO_TOOL_CALL_REPAIR_TTFT_TIMEOUT_MS;
delete process.env.KIRO_TOOL_CALL_REPAIR_STALL_TIMEOUT_MS;
});
afterEach(() => {
delete process.env.KIRO_TOOL_CALL_REPAIR_BUFFER_MAX_BYTES;
delete process.env.KIRO_TOOL_CALL_REPAIR_TTFT_TIMEOUT_MS;
delete process.env.KIRO_TOOL_CALL_REPAIR_STALL_TIMEOUT_MS;
});
describe("Kiro terminal integrity recovery", () => {
it("keeps semantic output private behind a heartbeat until clean EOF", async () => {
const upstream = controlledResponse([
frame("assistantResponseEvent", { content: "private until validated" })
]);
fetchMock.mockResolvedValueOnce(upstream.value);
const result = await execute();
const reader = result.response.body.getReader();
expect(new TextDecoder().decode((await reader.read()).value)).toBe(": kiro-validation\n\n");
let settled = false;
const semantic = reader.read().then((value) => {
settled = true;
return value;
});
await Promise.resolve();
expect(settled).toBe(false);
upstream.close();
expect(new TextDecoder().decode((await semantic).value)).toContain("private until validated");
await reader.cancel();
});
it("accepts CLI-compatible text and usage frames at clean EOF without messageStop", async () => {
fetchMock.mockResolvedValueOnce(response([
frame("assistantResponseEvent", { content: "Complete answer." }),
frame("meteringEvent", { usage: 2, unit: "credit" }),
frame("contextUsageEvent", { contextUsagePercentage: 10 })
]));
const body = await (await execute()).response.text();
expect(fetchMock).toHaveBeenCalledTimes(1);
expect(body).toContain("Complete answer.");
expect(body).toContain('"finish_reason":"stop"');
expect(body).toContain('"kiro_credits":2');
});
it("parses frames split across chunks and multiple frames in one chunk", async () => {
const first = frame("assistantResponseEvent", { content: "split " });
const second = frame("assistantResponseEvent", { content: "boundaries" });
const combined = concat([first, second]);
fetchMock.mockResolvedValueOnce(new Response(new ReadableStream({
start(controller) {
controller.enqueue(combined.slice(0, 9));
controller.enqueue(combined.slice(9, first.byteLength + 5));
controller.enqueue(combined.slice(first.byteLength + 5));
controller.close();
}
})));
const body = await (await execute()).response.text();
expect(body).toContain('"content":"split "');
expect(body).toContain('"content":"boundaries"');
expect(body).toContain('"finish_reason":"stop"');
});
it("accepts messageStop without semantic output as explicit completion", async () => {
fetchMock.mockResolvedValueOnce(response([frame("messageStopEvent", {})]));
const body = await (await execute()).response.text();
expect(fetchMock).toHaveBeenCalledTimes(1);
expect(body).toContain('"finish_reason":"stop"');
expect(body).not.toContain("kiro_missing_terminal");
});
it.each(["...", "…"])("repairs exact ellipsis final %s without leaking it", async (ellipsis) => {
fetchMock
.mockResolvedValueOnce(response([frame("assistantResponseEvent", { content: ellipsis })]))
.mockResolvedValueOnce(response([frame("assistantResponseEvent", { content: "Recovered answer." })]));
const body = await (await execute()).response.text();
expect(fetchMock).toHaveBeenCalledTimes(2);
expect(body).toContain("Recovered answer.");
expect(body).not.toContain(`"content":"${ellipsis}"`);
});
it.each([
"接下來我只再確認部署結果。",
"我會重新抓取最新日誌並確認結果。",
"目前證據顯示只在 **03:48:30–03:49:00 TPE** 出現少量 NonKA 504；主池 106/106、副池 50/50，且兩池都沒有重啟。最後補查 504 access log，確認 host／路徑與是否為集中流量。",
"Next I'll verify the deployment logs.",
"Let me check the remaining failures."
])("repairs conservative future-action final: %s", async (progress) => {
fetchMock
.mockResolvedValueOnce(response([frame("assistantResponseEvent", { content: progress })]))
.mockResolvedValueOnce(response([frame("assistantResponseEvent", { content: "Verification completed." })]));
const body = await (await execute()).response.text();
expect(fetchMock).toHaveBeenCalledTimes(2);
expect(body).toContain("Verification completed.");
expect(body).not.toContain(progress);
});
it.each([
"Working...",
"I'll check the logs. They show no errors and deployment succeeded.",
"Let me check: status is 200 and the checksum matches abc123.",
"我會檢查版本。版本是 1.2.3。",
"接下來請你先批准部署，我會等待你的確認。",
"已完成驗證，所有測試均通過。",
"目前證據顯示只有少量 504，且主副池均未重啟。",
"目前證據顯示只有少量 504。最後補查結果顯示沒有集中流量。",
"目前證據顯示只有少量 504。最後補查，結果顯示沒有集中流量。",
"目前證據顯示只有少量 504。最後補查：結果顯示沒有集中流量。",
"目前證據顯示只有少量 504。最後補查 504 access log，結果顯示沒有集中流量。",
"目前證據顯示只有少量 504。最後補查 504 access log，確認 host／路徑與有無集中流量：無集中流量。",
"目前證據顯示只有少量 504。最後補查 504 access log，確認 host／路徑與是否為集中流量（答案是否定的）。",
"目前證據顯示只有少量 504。最後補充兩點已確認的結果。",
"The verification is complete and all tests passed."
])("does not retry legitimate final: %s", async (finalText) => {
fetchMock.mockResolvedValueOnce(response([
frame("assistantResponseEvent", { content: finalText })
]));
const body = await (await execute()).response.text();
expect(fetchMock).toHaveBeenCalledTimes(1);
expect(body).toContain(finalText);
});
it("bounds incomplete-final repair to one retry", async () => {
fetchMock
.mockResolvedValueOnce(response([frame("assistantResponseEvent", { content: "..." })]))
.mockResolvedValueOnce(response([frame("assistantResponseEvent", { content: "…" })]));
const body = await (await execute()).response.text();
expect(fetchMock).toHaveBeenCalledTimes(2);
expect(body).toContain("kiro_ellipsis_retry_failed");
expect(body).not.toContain('"content":"..."');
});
it("repairs malformed wrapper tools without leaking the invalid call", async () => {
fetchMock
.mockResolvedValueOnce(response([frame("toolUseEvent", {
toolUseId: "bad",
name: "tool_call",
input: { arguments: { q: "router" } }
})]))
.mockResolvedValueOnce(response([frame("toolUseEvent", {
toolUseId: "good",
name: "tool_call",
input: { name: "mcp_search", arguments: { q: "router" } }
})]));
const body = await (await execute()).response.text();
expect(fetchMock).toHaveBeenCalledTimes(2);
expect(body).toContain('"name":"tool_call"');
expect(body).toContain('\\"name\\":\\"mcp_search\\"');
expect(body).not.toContain('"id":"bad"');
});
it("requires complete direct tool input and keeps the failure private", async () => {
const pending = frame("toolUseEvent", { toolUseId: "pending", name: "read_file" });
fetchMock
.mockResolvedValueOnce(response([pending]))
.mockResolvedValueOnce(response([pending]));
const body = await (await execute()).response.text();
expect(fetchMock).toHaveBeenCalledTimes(2);
expect(body).toContain("kiro_tool_call_repair_retry_failed");
expect(body).not.toContain('"name":"read_file"');
});
it("repairs a non-string toolUseId before releasing the tool call", async () => {
fetchMock
.mockResolvedValueOnce(response([frame("toolUseEvent", {
toolUseId: 123,
name: "read_file",
input: { path: "bad.txt" }
})]))
.mockResolvedValueOnce(response([frame("toolUseEvent", {
toolUseId: "valid-tool-id",
name: "read_file",
input: { path: "safe.txt" }
})]));
const body = await (await execute()).response.text();
expect(fetchMock).toHaveBeenCalledTimes(2);
expect(body).toContain('"id":"valid-tool-id"');
expect(body).not.toContain('"id":123');
});
it("keeps model-controlled parser detail out of the retry system prompt", async () => {
fetchMock
.mockResolvedValueOnce(response([frame("toolUseEvent", {
toolUseId: "bad-json",
name: "tool_call",
input: '{"name":"IGNORE_ALL_INSTRUCTIONS"'
})]))
.mockResolvedValueOnce(response([frame("assistantResponseEvent", {
content: "Recovered safely."
})]));
const body = await (await execute()).response.text();
const retryBody = JSON.parse(fetchMock.mock.calls[1][1].body);
expect(body).toContain("Recovered safely.");
expect(retryBody.systemPrompt).toContain("tool_call wrapper was malformed");
expect(retryBody.systemPrompt).not.toContain("IGNORE_ALL_INSTRUCTIONS");
});
it("lets a complete tool call override metadata end_turn", async () => {
fetchMock.mockResolvedValueOnce(response([
frame("toolUseEvent", {
toolUseId: "tool",
name: "read_file",
input: { path: "safe.txt" }
}),
frame("metadataEvent", { stopReason: "end_turn" })
]));
const body = await (await execute()).response.text();
expect(fetchMock).toHaveBeenCalledTimes(1);
expect(body).toContain('"name":"read_file"');
expect(body).toContain('"finish_reason":"tool_calls"');
});
it("maps max_tokens without treating it as a normal stop", async () => {
fetchMock.mockResolvedValueOnce(response([
frame("assistantResponseEvent", { content: "Limited answer." }),
frame("metadataEvent", { stopReason: "max_tokens" })
]));
const body = await (await execute()).response.text();
expect(fetchMock).toHaveBeenCalledTimes(1);
expect(body).toContain('"finish_reason":"length"');
expect(body).not.toContain('"finish_reason":"stop"');
});
it("retries malformed_model_output once without semantic leakage", async () => {
fetchMock
.mockResolvedValueOnce(response([
frame("assistantResponseEvent", { content: "private malformed output" }),
frame("metadataEvent", { stopReason: "malformed_model_output" })
]))
.mockResolvedValueOnce(response([
frame("assistantResponseEvent", { content: "Recovered protocol output." })
]));
const body = await (await execute()).response.text();
expect(fetchMock).toHaveBeenCalledTimes(2);
expect(body).toContain("Recovered protocol output.");
expect(body).not.toContain("private malformed output");
});
it.each([
["cancelled", "kiro_terminal_incomplete"],
["pause_turn", "kiro_terminal_incomplete"],
["content_filtered", "kiro_terminal_refusal"],
["novel_reason", "kiro_unknown_stop_reason"]
])("fails closed for stop reason %s", async (stopReason, code) => {
fetchMock.mockResolvedValueOnce(response([
frame("assistantResponseEvent", { content: `private-${stopReason}` }),
frame("metadataEvent", { stopReason })
]));
const body = await (await execute()).response.text();
expect(fetchMock).toHaveBeenCalledTimes(1);
expect(body).toContain(code);
expect(body).not.toContain(`private-${stopReason}`);
expect(body).not.toContain('"finish_reason":"stop"');
});
it.each([
[
frame("messageStopEvent", { stopReason: "content_filtered" }),
frame("metadataEvent", { stopReason: "end_turn" })
],
[
frame("metadataEvent", { stopReason: "end_turn" }),
frame("messageStopEvent", { stopReason: "content_filtered" })
]
])("preserves the most restrictive conflicting stop reason", async (...stopFrames) => {
fetchMock.mockResolvedValueOnce(response([
frame("assistantResponseEvent", { content: "private filtered output" }),
...stopFrames
]));
const body = await (await execute()).response.text();
expect(fetchMock).toHaveBeenCalledTimes(1);
expect(body).toContain("kiro_terminal_refusal");
expect(body).not.toContain("private filtered output");
});
it("prefers a non-retryable terminal reason over an earlier retryable reason", async () => {
fetchMock.mockResolvedValueOnce(response([
frame("assistantResponseEvent", { content: "private malformed output" }),
frame("metadataEvent", { stopReason: "malformed_model_output" }),
frame("messageStopEvent", { stopReason: "cancelled" })
]));
const body = await (await execute()).response.text();
expect(fetchMock).toHaveBeenCalledTimes(1);
expect(body).toContain("kiro_terminal_incomplete");
expect(body).toContain('"stop_reason":"cancelled"');
expect(body).not.toContain("private malformed output");
});
it("preserves an authoritative refusal returned by the bounded retry", async () => {
fetchMock
.mockResolvedValueOnce(response([]))
.mockResolvedValueOnce(response([
frame("assistantResponseEvent", { content: "private filtered retry" }),
frame("metadataEvent", { stopReason: "content_filtered" })
]));
const body = await (await execute()).response.text();
expect(fetchMock).toHaveBeenCalledTimes(2);
expect(body).toContain("kiro_terminal_refusal");
expect(body).not.toContain("kiro_missing_terminal_retry_failed");
expect(body).not.toContain("private filtered retry");
});
it.each([
["max_tokens", "kiro_terminal_incomplete"],
["cancelled", "kiro_terminal_incomplete"],
["content_filtered", "kiro_terminal_refusal"],
["novel_reason", "kiro_unknown_stop_reason"]
])("does not let a valid tool override failure stop reason %s", async (stopReason, code) => {
fetchMock.mockResolvedValueOnce(response([
frame("toolUseEvent", {
toolUseId: "blocked-tool",
name: "read_file",
input: { path: "secret.txt" }
}),
frame("metadataEvent", { stopReason })
]));
const body = await (await execute()).response.text();
expect(fetchMock).toHaveBeenCalledTimes(1);
expect(body).toContain(code);
expect(body).not.toContain('"name":"read_file"');
});
it.each(["content_filtered", "cancelled", "max_tokens"])(
"classifies failure %s before validating a malformed deferred tool",
async (stopReason) => {
fetchMock.mockResolvedValueOnce(response([
frame("toolUseEvent", { toolUseId: "bad-tool", name: "read_file" }),
frame("metadataEvent", { stopReason })
]));
const body = await (await execute()).response.text();
expect(fetchMock).toHaveBeenCalledTimes(1);
expect(body).toContain(stopReason === "content_filtered"
? "kiro_terminal_refusal"
: "kiro_terminal_incomplete");
expect(body).not.toContain("kiro_tool_call_repair_retry_failed");
expect(body).not.toContain('"name":"read_file"');
}
);
it.each([
["content_filtered", [frame("toolUseEvent", {
toolUseId: 123,
name: "read_file",
input: { path: "bad.txt" }
})], "kiro_terminal_refusal"],
["cancelled", [frame("toolUseEvent", {
toolUseId: "missing-name",
input: { path: "bad.txt" }
})], "kiro_terminal_incomplete"],
["max_tokens", [
frame("toolUseEvent", { toolUseId: "changing", name: "read_file" }),
frame("toolUseEvent", { toolUseId: "changing", name: "write_file" })
], "kiro_terminal_incomplete"]
])("continues past eager tool-shape errors to authoritative stop %s", async (stopReason, toolFrames, code) => {
fetchMock.mockResolvedValueOnce(response([
...toolFrames,
frame("metadataEvent", { stopReason })
]));
const body = await (await execute()).response.text();
expect(fetchMock).toHaveBeenCalledTimes(1);
expect(body).toContain(code);
expect(body).not.toContain("kiro_tool_call_repair_retry_failed");
expect(body).not.toContain('"tool_calls"');
});
it("retries a TTFT timeout once while preserving cancellation semantics", async () => {
process.env.KIRO_TOOL_CALL_REPAIR_TTFT_TIMEOUT_MS = "1";
fetchMock
.mockResolvedValueOnce(controlledResponse().value)
.mockResolvedValueOnce(response([
frame("assistantResponseEvent", { content: "Recovered after timeout." })
]));
const body = await (await execute()).response.text();
expect(fetchMock).toHaveBeenCalledTimes(2);
expect(body).toContain("Recovered after timeout.");
});
it("treats validated non-semantic frames as watchdog activity", async () => {
process.env.KIRO_TOOL_CALL_REPAIR_TTFT_TIMEOUT_MS = "30";
process.env.KIRO_TOOL_CALL_REPAIR_STALL_TIMEOUT_MS = "30";
const upstream = controlledResponse();
fetchMock.mockResolvedValueOnce(upstream.value);
setTimeout(() => upstream.enqueue(frame("meteringEvent", { usage: 1 })), 20);
setTimeout(() => upstream.enqueue(frame("contextUsageEvent", { contextUsagePercentage: 5 })), 40);
setTimeout(() => {
upstream.enqueue(frame("assistantResponseEvent", { content: "Completed after active frames." }));
upstream.close();
}, 60);
const body = await (await execute()).response.text();
expect(fetchMock).toHaveBeenCalledTimes(1);
expect(body).toContain("Completed after active frames.");
});
it("retries a response-body read failure once", async () => {
fetchMock
.mockResolvedValueOnce(new Response(new ReadableStream({
start(controller) {
controller.error(new Error("socket reset"));
}
})))
.mockResolvedValueOnce(response([
frame("assistantResponseEvent", { content: "Recovered after read failure." })
]));
const body = await (await execute()).response.text();
expect(fetchMock).toHaveBeenCalledTimes(2);
expect(body).toContain("Recovered after read failure.");
expect(body).not.toContain("socket reset");
});
it.each([
["message CRC", () => {
const corrupt = frame("assistantResponseEvent", { content: "corrupt CRC" });
corrupt[corrupt.byteLength - 1] ^= 0xff;
return [corrupt];
}],
["prelude CRC", () => {
const corrupt = frame("assistantResponseEvent", { content: "corrupt prelude" });
corrupt[8] ^= 0xff;
return [corrupt];
}],
["truncated frame", () => {
const truncated = frame("assistantResponseEvent", { content: "truncated" });
return [truncated.slice(0, -3)];
}],
["out-of-bounds headers", () => {
const corrupt = frame("assistantResponseEvent", { content: "bad headers" });
new DataView(corrupt.buffer).setUint32(4, corrupt.byteLength - 15, false);
return [checksum(corrupt)];
}],
["duplicate headers", () => [
frameFromEntries([
[":event-type", "assistantResponseEvent"],
[":event-type", "metadataEvent"]
], { content: "duplicate" })
]]
])("retries %s and releases only the valid attempt", async (_name, invalidFrames) => {
fetchMock
.mockResolvedValueOnce(response([
frame("assistantResponseEvent", { content: "must stay private" }),
...invalidFrames()
]))
.mockResolvedValueOnce(response([
frame("assistantResponseEvent", { content: "Recovered after validation." })
]));
const body = await (await execute()).response.text();
expect(fetchMock).toHaveBeenCalledTimes(2);
expect(body).toContain("Recovered after validation.");
expect(body).not.toContain("must stay private");
});
it("reports corrupt-frame provenance when the bounded retry also fails", async () => {
const corruptFrame = () => {
const corrupt = frame("assistantResponseEvent", { content: "corrupt" });
corrupt[corrupt.byteLength - 1] ^= 0xff;
return corrupt;
};
fetchMock
.mockResolvedValueOnce(response([corruptFrame()]))
.mockResolvedValueOnce(response([corruptFrame()]));
const body = await (await execute()).response.text();
expect(body).toContain("kiro_missing_terminal_retry_failed");
expect(body).toContain('"terminal_provenance":"corrupt_eventstream_frame"');
expect(body).toContain('"transport_state":"corrupt_frame"');
});
it("caps diagnostic event-type cardinality", async () => {
let terminal;
const executor = new KiroExecutor();
const frames = Array.from({ length: 100 }, (_, index) =>
frame(`unknownEvent${index}`, { index })
);
frames.push(frame("assistantResponseEvent", { content: "done" }));
const transformed = executor.transformEventStreamToSSE(
response(frames),
"kr/claude-opus-4.8",
{ onTerminalState: (value) => { terminal = value; } }
);
await transformed.text();
expect(terminal.event_counts).toEqual({
other: 100,
assistantResponseEvent: 1
});
});
it("rejects a raw chunk before concatenating beyond the protocol bound", async () => {
let terminal;
const executor = new KiroExecutor();
const transformed = executor.transformEventStreamToSSE(
response([new Uint8Array(65)]),
"kr/claude-opus-4.8",
{
maxRawBytes: 64,
onTerminalState: (value) => { terminal = value; }
}
);
const body = await transformed.text();
expect(body).toContain("buffered bytes exceed the protocol bound");
expect(terminal.terminal_provenance).toBe("corrupt_eventstream_frame");
});
it.each(["error", "exception"])("propagates EventStream %s without retry or leakage", async (messageType) => {
fetchMock.mockResolvedValueOnce(response([
frame("assistantResponseEvent", { content: "must stay private" }),
frameFromEntries([
[":message-type", messageType],
...(messageType === "exception" ? [[":exception-type", "InternalServerException"]] : [])
], { message: "upstream failed" })
]));
const body = await (await execute()).response.text();
expect(fetchMock).toHaveBeenCalledTimes(1);
expect(body).toContain("kiro_upstream_eventstream_error");
expect(body).toContain("upstream failed");
expect(body).not.toContain("must stay private");
});
it("surfaces retry HTTP failures as SSE after heartbeat commits headers", async () => {
fetchMock
.mockResolvedValueOnce(response([]))
.mockResolvedValueOnce(new Response("unauthorized", {
status: 401,
statusText: "Unauthorized"
}));
const result = await execute();
const body = await result.response.text();
expect(result.response.status).toBe(200);
expect(body).toContain("kiro_integrity_retry_upstream_error");
expect(body).toContain("unauthorized");
});
it("bounds the retry HTTP error body", async () => {
fetchMock
.mockResolvedValueOnce(response([]))
.mockResolvedValueOnce(new Response(`error-start-${"x".repeat(10_000)}-error-tail`, {
status: 401,
statusText: "Unauthorized"
}));
const body = await (await execute()).response.text();
expect(body).toContain("error-start-");
expect(body).not.toContain("error-tail");
expect(body.length).toBeLessThan(5000);
});
it("propagates cancellation while validation is waiting for EOF", async () => {
const upstream = controlledResponse([
frame("assistantResponseEvent", { content: "waiting" })
]);
fetchMock.mockResolvedValueOnce(upstream.value);
const abort = new AbortController();
const result = await execute(new KiroExecutor(), { signal: abort.signal });
const reader = result.response.body.getReader();
await reader.read();
abort.abort("client cancelled");
await expect(reader.read()).rejects.toMatchObject({ name: "AbortError" });
});
it("fails safely when the private gate exceeds its configured bound", async () => {
process.env.KIRO_TOOL_CALL_REPAIR_BUFFER_MAX_BYTES = "8";
fetchMock.mockResolvedValueOnce(response([
frame("assistantResponseEvent", { content: "larger than eight bytes" })
]));
const body = await (await execute()).response.text();
expect(fetchMock).toHaveBeenCalledTimes(1);
expect(body).toContain("integrity buffer exceeded");
expect(body).not.toContain("larger than eight bytes");
});
it("counts deferred tool fragments against the private memory bound", async () => {
process.env.KIRO_TOOL_CALL_REPAIR_BUFFER_MAX_BYTES = "128";
fetchMock.mockResolvedValueOnce(response([
frame("toolUseEvent", {
toolUseId: "large-tool",
name: "read_file",
input: { path: "x".repeat(200) }
})
]));
const body = await (await execute()).response.text();
expect(fetchMock).toHaveBeenCalledTimes(1);
expect(body).toContain("kiro_integrity_buffer_exceeded");
expect(body).not.toContain('"name":"read_file"');
});
});
tests/unit/kiro-thinking-strip.test.js
+26 -6
View File
@@ -32,10 +32,23 @@ function createMockFrame(eventType, payloadObj) {
offset += headerValueBytes.length;
buffer.set(payloadBytes, offset);
view.setUint32(8, crc32(buffer.subarray(0, 8)), false);
view.setUint32(totalLength - 4, crc32(buffer.subarray(0, totalLength - 4)), false);
return buffer;
}
function crc32(bytes) {
let crc = 0xffffffff;
for (const byte of bytes) {
crc ^= byte;
for (let bit = 0; bit < 8; bit++) {
crc = (crc >>> 1) ^ ((crc & 1) ? 0xedb88320 : 0);
}
}
return (crc ^ 0xffffffff) >>> 0;
}
async function readAllSSE(stream) {
const reader = stream.getReader();
const decoder = new TextDecoder();
@@ -130,14 +143,16 @@ describe("KiroExecutor thinking tag stripping", () => {
expect(contentChunks.length).toBe(0);
});
it("emits a terminal chunk at messageStop before the upstream stream closes", async () => {
it("waits for clean EOF before emitting stop after messageStop", async () => {
const executor = new KiroExecutor();
const f1 = createMockFrame("assistantResponseEvent", { content: "OK" });
const f2 = createMockFrame("messageStopEvent", {});
let upstreamController;
const readableStream = new ReadableStream({
start(controller) {
upstreamController = controller;
controller.enqueue(f1);
controller.enqueue(f2);
}
@@ -147,11 +162,16 @@ describe("KiroExecutor thinking tag stripping", () => {
const reader = transformedResponse.body.getReader();
const decoder = new TextDecoder();
let output = "";
for (let i = 0; i < 4 && !output.includes("\"finish_reason\":\"stop\""); i++) {
const { value } = await readNextWithTimeout(reader);
output += decoder.decode(value, { stream: true });
const { value } = await readNextWithTimeout(reader);
output += decoder.decode(value, { stream: true });
expect(output).not.toContain("\"finish_reason\":\"stop\"");
upstreamController.close();
while (!output.includes("\"finish_reason\":\"stop\"")) {
const { value: nextValue, done } = await readNextWithTimeout(reader);
if (done) break;
output += decoder.decode(nextValue, { stream: true });
}
await reader.cancel();
expect(output).toContain("\"finish_reason\":\"stop\"");
});
tests/unit/openai-to-kiro.test.js
+96
View File
@@ -316,6 +316,102 @@ describe("openaiToKiroRequest", () => {
});
});
it.each([
["high", "gpt-5.6-sol"],
["medium", "kiro/gpt-5.6-terra"],
["low", "gpt-5.6-luna"],
])("maps GPT-5.6 reasoning.effort %s without legacy prompt tags", (effort, model) => {
const body = {
reasoning: { effort },
messages: [{ role: "user", content: "Use the requested effort" }]
};
const result = openaiToKiroRequest(model, body, true, {});
expect(result.additionalModelRequestFields).toEqual({
reasoning: { effort },
});
expect(systemPromptOf(result)).not.toContain("<thinking_mode>");
expect(systemPromptOf(result)).not.toContain("<max_thinking_length>");
expect(contentOf(result)).not.toContain("<thinking_mode>");
expect(contentOf(result)).not.toContain("<max_thinking_length>");
});
it.each([
["xhigh", "gpt-5.6-terra", "xhigh"],
["max", "gpt-5.6-sol", "xhigh"],
])("preserves GPT-5.6 effort %s as supported wire effort %s", (effort, model, wireEffort) => {
const body = {
reasoning: { effort },
messages: [{ role: "user", content: "Use extended effort" }]
};
const result = openaiToKiroRequest(model, body, true, {});
expect(result.additionalModelRequestFields).toEqual({
reasoning: { effort: wireEffort },
});
expect(systemPromptOf(result)).not.toContain("<thinking_mode>");
expect(systemPromptOf(result)).not.toContain("<max_thinking_length>");
});
it("omits GPT-5.6 effort fields and legacy prompt tags when effort is absent", () => {
const body = {
messages: [{ role: "user", content: "No explicit reasoning effort" }]
};
const result = openaiToKiroRequest("gpt-5.6-sol", body, true, {});
expect(result.additionalModelRequestFields).toBeUndefined();
expect(systemPromptOf(result)).not.toContain("<thinking_mode>");
expect(systemPromptOf(result)).not.toContain("<max_thinking_length>");
});
it.each(["auto", "minimal", "ultra"])(
"keeps the legacy thinking fallback for unsupported GPT-5.6 effort %s",
(effort) => {
const body = {
reasoning: { effort },
messages: [{ role: "user", content: "Use legacy thinking" }]
};
const result = openaiToKiroRequest("gpt-5.6-luna", body, true, {});
expect(result.additionalModelRequestFields).toBeUndefined();
expect(systemPromptOf(result)).toContain("<thinking_mode>enabled</thinking_mode>");
expect(systemPromptOf(result)).toContain("<max_thinking_length>");
}
);
it.each(["none", "off", "disabled"])(
"keeps GPT-5.6 reasoning intentionally disabled for effort %s",
(effort) => {
const body = {
reasoning: { effort },
messages: [{ role: "user", content: "Do not reason" }]
};
const result = openaiToKiroRequest("gpt-5.6-luna", body, true, {});
expect(result.additionalModelRequestFields).toBeUndefined();
expect(systemPromptOf(result)).not.toContain("<thinking_mode>");
expect(systemPromptOf(result)).not.toContain("<max_thinking_length>");
}
);
it("keeps the thinking-alias fallback when GPT effort is blank", () => {
const body = {
reasoning: { effort: "" },
messages: [{ role: "user", content: "Use the thinking alias" }]
};
const result = openaiToKiroRequest("gpt-5.6-sol-thinking", body, true, {});
expect(result.additionalModelRequestFields).toBeUndefined();
expect(systemPromptOf(result)).toContain("<thinking_mode>enabled</thinking_mode>");
expect(systemPromptOf(result)).toContain("<max_thinking_length>");
});
it("does not send additionalModelRequestFields for legacy Kiro model ids", () => {
const body = {
reasoning_effort: "high",