Most teams end up with one account per model vendor, each with its own key, invoice and quirks. UnoRouter collapses that into a single OpenAI compatible endpoint: the same request reaches GPT, Claude, Gemini, DeepSeek, Qwen and dozens of other model families, and a provider outage is routed around automatically. It plugs into Claude Code, Cline, SillyTavern or your own scripts with a base URL change. The whole stack is open source and the chat keeps history in your browser. Free tier first, then pay per use with no subscription.
UnoRouter provides an API for building apps and running agents, with OpenAI-compatible endpoints, API documentation, and integrations for coding agents and CLI tools.
One API key connects users to more than 280 AI models from multiple providers through a unified endpoint.
OpenAI compatibility
OpenAI-compatible endpoints let users switch providers without changing their code.
Provider failover
Automatic provider failover can route requests to another provider when one is unavailable.
Built-in character chat
Users can chat on the site with free models, choose characters, or import their own without setting up an account or API key.
Browser-stored conversations
The site says character-chat conversations stay in the user's browser, and users can bring their own OpenAI-compatible endpoint and key.
Who it’s for
App developers and people using coding agents, CLI tools, or compatible chat clients who want one API key for multiple AI models. It also serves people who want to chat with characters on the site without setup.
What it costs
Free models cost 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 the credit never expires; no credit card is needed to start.
