What it does
- One API key gives access to 287+ models from 57+ providers, including GPT, Claude, Gemini, DeepSeek and GLM.
- Models ending in :free cost nothing and are ready to use instantly, with a limit of about one request per minute per model.
- Endpoints are fully OpenAI compatible, so existing code works without changes.
- Automatic provider failover reroutes requests to another upstream when one is busy.
- The built-in chat client supports personas, presets, lorebooks, and SillyTavern card import, with no signup or setup.
- Paid models are billed per use from a $1 top-up, and credit never expires.
Pricing
All models ending in :free are free with rate limits and no credit card required. Paid models are pay as you go from a $1 top-up, charged per token, and credit never expires.
Good to know
- Works on Web, Linux, macOS, and Windows.
- Chats in the chat app are stored in your browser, not on the site's servers, and all code is open source.
- Some providers keep or train on inputs; you see a notice or give consent once per provider before requests reach them.
Questions & answers
- Is UnoRouter free?
- Yes. Every model ending in :free costs nothing, with a limit of about one request per minute per model. Paid models are pay as you go from a $1 top up, and that credit never expires. No credit card is needed to start.
- Is UnoRouter safe to use?
- Chats in the UnoRouter chat app are stored in your browser, not on our servers. Some providers keep or train on inputs; for those you see a notice or give consent once per provider before any request reaches them. All of the code is open source.
- Is UnoRouter down?
- Usually not. When a free model stops answering, every provider for that model is busy for a few minutes. Switch to another free model, or check the status page, which shows each model live.
In their words
One API key for every AI model
Free models, no credit card