Qwen Free API — Go Bridge for chat.qwen.ai

Qwen Free API is my Go entry in the free-API bridge family: it speaks to chat.qwen.ai (Qwen3.8-Max) and exposes it behind OpenAI- and Anthropic-compatible endpoints, so any tool that expects those protocols can use Qwen for free.

Design

• Single file, zero browser — like GLM Free API, no headless Chrome; pure HTTP session handling in Go.
• Dual protocol — /v1/chat/completions (OpenAI) and /v1/messages (Anthropic) from one binary.
• SSE streaming passthrough — chunks relay as they arrive, keeping token-by-token latency honest.
• Session hygiene — guest-token acquisition, rotation and retry baked into the request loop.

./qwen-free-api --port 8005
curl localhost:8005/v1/chat/completions \
  -d '{"model":"qwen-max","messages":[{"role":"user","content":"سلام"}],"stream":true}'

Same philosophy as the rest of the family: the bridge must never be the thing that is down — timeouts, retries and actionable errors, not raw tracebacks. Point OmniRouter at it and Qwen becomes one more upstream in the pool. MIT licensed.