UnoRouter

UnoRouter is a unified AI gateway that exposes a fully OpenAI-compatible API, letting you access many models through a single endpoint and one key. With an OpenAI-style client, you can keep your existing request flow while selecting from a wide set of model names.

In addition to standard API routing, UnoRouter supports multiple client types, including coding agents, character & fiction clients, general chat clients, and CLI-style tools. It also includes a built-in chat experience for trying models directly in the browser.

UnoRouter is designed with resilience in mind: it includes automatic provider failover so requests can continue when upstream providers are unavailable. For character chat workflows, you can also use a private/BYOK mode where conversation data stays in your browser and you bring your own OpenAI-compatible endpoint and key for browser-fired requests.

Is UnoRouter compatible with OpenAI-style clients?

Yes. You can point an OpenAI-compatible chat/completions client at UnoRouter and use UnoRouter’s API key so your requests route to the selected model.

How does provider failover work in UnoRouter?

UnoRouter is built to route requests across providers and keep going when an upstream provider isn’t available, reducing downtime during outages.

Can UnoRouter be used with character or fiction chat tools?

Yes. UnoRouter can be used with character & fiction clients such as SillyTavern and others that support OpenAI-compatible endpoints.

What’s the difference between built-in chat and private/BYOK chat?

Built-in chat lets you use free models directly in the browser. Private/BYOK chat keeps conversation data in your browser and lets you bring your own OpenAI-compatible endpoint and key.

Do I need to change my existing integration logic to use UnoRouter?

Typically not. If your app already speaks OpenAI-compatible APIs, you mainly swap to UnoRouter’s endpoint and key while keeping the same request format.