🚀 UnQRoute — The Free AI Gateway

Never stop coding. Connect every AI tool to 237 providers — 90+ free — through one endpoint.

Plug Claude Code, Codex, Cursor, Cline, Copilot & Antigravity into FREE Claude / GPT / Gemini. Auto-fallback.

Includes Z.ai Web (GLM-5.2) integration with token-pool architecture for 99.999%+ success rate.

RTK + Caveman compression saves 15–95% tokens. Never hit limits.

Forked from OmniRoute. Rebranded as UnQRoute with custom Z.ai Web (GLM-5.2) provider integration.

📦 Install: npm install -g unqroute · Then run: unqroute

🧩 Available CLI commands

unqroute                    # Start the server (alias: unqroute serve)
unqroute serve              # Start server with options
unqroute setup              # Interactive setup wizard
unqroute status             # Show server status
unqroute providers          # List providers
unqroute --help             # Show all commands

🌐 Quick API usage

# Once server is running on http://localhost:20128
curl http://localhost:20128/v1/chat/completions \
  -H "Authorization: Bearer <your-api-key>" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "zai/glm-5.2",
    "messages": [{"role": "user", "content": "Hello!"}]
  }'

What's New in UnQRoute

This fork adds a Z.ai Web (GLM-5.2) provider (zai-web / alias zai) that uses chat.z.ai's reverse-engineered internal API to expose GLM-5.2 as an OpenAI-compatible LLM via user JWT cookies.

Architecture:

• Token pool of N user JWTs stored in provider_connections (provider="zai-web")
• On each request, executor picks the best healthy token, signs the request with HMAC-SHA256 (hard-coded secret cracked from chat.z.ai's JS bundle), and POSTs to /api/v2/chat/completions
• If response contains FRONTEND_CAPTCHA_REQUIRED (Aliyun slider captcha, triggered per-request by server-side risk scoring), the token is put in 30-min cooldown and the next token is tried
• With N tokens, success rate = 1 - 0.2^N (assuming ~80% per-token success)
• For 99.999% success, import 25+ tokens via dashboard bulk-import

To use:

1. Create 25+ Z.ai accounts at chat.z.ai (email/Google signup)
2. For each account: login → DevTools → Application → Cookies → copy token value
3. Bulk-import all tokens via UnQRoute dashboard → Providers → "Z.ai Web (GLM-5.2)" → bulk-import
4. Call the API with model: "zai/glm-5.2"

Backward Compatibility

• ~/.omniroute/ data directory is still detected and used if it exists (existing OmniRoute users keep their data on upgrade)
• The omniroute CLI command is still installed as an alias for unqroute
• All OMNIROUTE_* environment variables continue to work

License

MIT — same as upstream OmniRoute