feat(vercel-ai-gateway): support embeddings, images and credit usage

Extend Vercel AI Gateway beyond chat: add OpenAI-compatible embeddings
and image generation endpoints, credit balance fetch on the usage
dashboard, retry on 429, and models catalog fetcher.

Thinking/reasoning mapping is omitted pending a project-wide refactor.

Co-authored-by: Ngô Tấn Tài <tantai@newnol.io.vn>
Co-authored-by: Cursor <cursoragent@cursor.com>

open-sse/services/usage.js

@@ -30,6 +30,11 @@ const MINIMAX_USAGE_URLS = {
   ],
 };
 
+// Vercel AI Gateway credits endpoint
+// Returns { balance: "95.50", total_used: "4.50" } (USD as decimal strings).
+// Docs: https://vercel.com/docs/ai-gateway/usage
+const VERCEL_AI_GATEWAY_CREDITS_URL = "https://ai-gateway.vercel.sh/v1/credits";
+
 // Antigravity API config (from Quotio)
 const ANTIGRAVITY_CONFIG = {
   quotaApiUrl: "https://cloudcode-pa.googleapis.com/v1internal:fetchAvailableModels",
@@ -92,6 +97,8 @@ export async function getUsageForProvider(connection, proxyOptions = null) {
     case "minimax":
     case "minimax-cn":
       return await getMiniMaxUsage(apiKey, provider, proxyOptions);
+    case "vercel-ai-gateway":
+      return await getVercelAiGatewayUsage(apiKey, proxyOptions);
     default:
       return { message: `Usage API not implemented for ${provider}` };
   }
@@ -1203,6 +1210,89 @@ async function getMiniMaxUsage(apiKey, provider, proxyOptions = null) {
   return { message: lastErrorMessage ? `MiniMax connected. Unable to fetch usage: ${lastErrorMessage}` : "MiniMax connected. Unable to fetch usage." };
 }
 
+
+/**
+ * Vercel AI Gateway usage — credit balance for the API key
+ *
+ * Calls GET /v1/credits which returns:
+ *   { "balance": "95.50", "total_used": "4.50" }   (USD as decimal strings)
+ *
+ * We surface this as a single "Balance ($)" quota row so the existing
+ * QuotaTable / progress-bar UI can render it. used = total_used,
+ * total = balance + total_used (the original credit allotment), so the
+ * remaining percentage equals balance / total.
+ *
+ * Docs: https://vercel.com/docs/ai-gateway/usage
+ */
+async function getVercelAiGatewayUsage(apiKey, proxyOptions = null) {
+  if (!apiKey) {
+    return { message: "Vercel AI Gateway API key not available." };
+  }
+
+  try {
+    const response = await proxyAwareFetch(VERCEL_AI_GATEWAY_CREDITS_URL, {
+      method: "GET",
+      headers: {
+        Authorization: `Bearer ${apiKey}`,
+        Accept: "application/json",
+      },
+    }, proxyOptions);
+
+    if (response.status === 401 || response.status === 403) {
+      return { message: "Vercel AI Gateway API key invalid or expired." };
+    }
+
+    if (!response.ok) {
+      const errorText = await response.text().catch(() => "");
+      const trimmed = errorText ? `: ${errorText.slice(0, 200)}` : "";
+      return { message: `Vercel AI Gateway credits API error (${response.status})${trimmed}` };
+    }
+
+    const data = await response.json();
+
+    // Vercel returns numeric strings; coerce safely.
+    const balance = Number(data?.balance) || 0;
+    const totalUsed = Number(data?.total_used) || 0;
+
+    // Vercel gives $5/month free credit. The API doesn't return the
+    // monthly allocation so we use the known constant as the denominator.
+    const MONTHLY_CREDIT = 5;
+    const remainingPercentage = (balance / MONTHLY_CREDIT) * 100;
+
+    if (balance <= 0 && totalUsed <= 0) {
+      return {
+        plan: "Pay-as-you-go",
+        message: "Vercel AI Gateway connected. No credit allocation found (BYOK or unfunded account).",
+        quotas: {},
+      };
+    }
+
+    // "Used (USD)": how much has been spent this month (no fixed cap → unlimited).
+    // "Remaining (USD)": balance remaining out of the $5 monthly allocation.
+    return {
+      plan: "Pay-as-you-go",
+      quotas: {
+        "Used (USD)": {
+          used: totalUsed,
+          total: 0,
+          remaining: 0,
+          remainingPercentage: 100,
+          unlimited: true,
+        },
+        "Remaining (USD)": {
+          used: balance,
+          total: MONTHLY_CREDIT,
+          remaining: balance,
+          remainingPercentage,
+          unlimited: false,
+        },
+      },
+    };
+  } catch (error) {
+    return { message: `Vercel AI Gateway error: ${error.message}` };
+  }
+}
+
 async function getQoderUsage(accessToken, proxyOptions = null) {
   if (!accessToken) {
     return { message: "Qoder usage unavailable: no access token" };