// app/api/chat/route.ts - Add basic rate limiting
import { Ratelimit } from "@upstash/ratelimit";
import { Redis } from "@upstash/redis";

const ratelimit = new Ratelimit({
  redis: Redis.fromEnv(),
  limiter: Ratelimit.slidingWindow(10, "1 m"), // 10 requests per minute
});

export async function POST(request: Request) {
  const ip = request.headers.get("x-forwarded-for") ?? "anonymous";
  const { success } = await ratelimit.limit(ip);

  if (!success) {
    return new Response("Rate limit exceeded", { status: 429 });
  }
  // ... rest of handler
}

// lib/ai/errors.ts
export class AIProviderError extends Error {
  constructor(
    public provider: AIProvider,
    message: string,
    public retryable: boolean = false,
  ) {
    super(message);
  }
}

// app/api/chat/route.ts - Add retry logic
const MAX_RETRIES = 2;
for (let attempt = 0; attempt <= MAX_RETRIES; attempt++) {
  try {
    return await streamResponse(messages);
  } catch (error) {
    if (attempt === MAX_RETRIES || !isRetryable(error)) throw error;
    await delay(1000 * attempt);
  }
}

// Track key metrics
const startTime = performance.now();
const result = streamText({ ... });
const ttft = performance.now() - startTime;

// Log to Vercel Analytics or custom endpoint
trackMetric("ai_ttft", ttft);
trackMetric("ai_provider", provider);
trackMetric("ai_success", !error);

# Rate limiting (Upstash Redis)
UPSTASH_REDIS_REST_URL=https://xxx.upstash.io
UPSTASH_REDIS_REST_TOKEN=xxx

# Budget alerts (optional)
AI_MONTHLY_BUDGET_USD=100
AI_ALERT_EMAIL=admin@example.com

┌─────────────────────────────────────────────────────────────────┐
│  app/page.tsx                                                   │
│  └── useChatProvider (feature flag switch)                     │
│       ├── NEXT_PUBLIC_USE_VERCEL_AI=true                       │
│       │   └── useVercelChat → /api/chat → Claude/GPT           │
│       │       ├── ✅ Real streaming                            │
│       │       ├── ✅ Abort on unmount                          │
│       │       ├── ✅ Memoized messages                         │
│       │       └── ✅ Fixed regenerate                          │
│       │                                                         │
│       └── NEXT_PUBLIC_USE_VERCEL_AI=false (default)            │
│           └── useAiChat → /api/ai/query → vLLM                 │
│               ├── ⚠️ Fake streaming (waits for full response)  │
│               ├── ✅ AbortController cleanup                   │
│               └── ✅ Fixed regenerate                          │
└─────────────────────────────────────────────────────────────────┘