What is ToRouter
ToRouter is a unified AI API gateway — one key, every model, one bill.
ToRouter is a unified AI API gateway. You get one API key and one base URL, and we route your traffic to OpenAI, Anthropic, Google, DeepSeek, Qwen, Moonshot, MiniMax, OpenRouter and other upstream providers behind the scenes.
It is OpenAI-, Anthropic- and Gemini-protocol compatible, so the SDK you already use keeps working — you only change base_url and the API key.
Why ToRouter
One key for every model
Stop juggling six accounts. Subscribe in the console, get one ToRouter key, and call models from the catalog.
Drop-in compatible
The OpenAI / Anthropic / Gemini SDKs work without code changes — only base_url and key change.
Pay-per-token
No monthly minimum. Top up once, spend as you go. Per-model pricing is visible up front.
Built-in observability
Per-key quota, per-model spend, per-channel uptime — all in one dashboard.
What you can do with ToRouter
- Call OpenAI GPT-5 / GPT-4 / o-series, Anthropic Claude (Opus / Sonnet / Haiku), Google Gemini, DeepSeek, Qwen, GLM, Kimi, MiniMax and more from a single endpoint
- Use the OpenAI Chat Completions / Responses / Embeddings / Images API
- Use the native Anthropic
/v1/messagesand Gemini/v1betaAPIs - Plug into Cursor, Cline, Claude Code, LangChain, Dify, n8n and any OpenAI-compatible client
- Set per-key IP whitelist (CIDR), rate limits, model whitelist, expiration
- Get failover across channels when an upstream provider has issues
Where to start
5-minute quickstart
Sign up, top up, create a key, make your first request.
Core concepts
Account, channel, model, key, quota — the mental model in one page.
Base URL & authentication
The endpoint table for OpenAI / Anthropic / Gemini SDKs.
SDK integrations
Python, Node, Cursor, Cline, Claude Code and more.
The console lives at portal.torouter.ai. API traffic uses the same host.