An open source, OpenAI compatible AI gateway: one API key for models from many providers, plus a built in chat client.

Which models can be used?

The site lists models from OpenAI, Anthropic, Google, DeepSeek, Moonshot, Zhipu, Qwen and other providers.

Are there free models?

Yes. Free models use the :free suffix, route only to free providers and do not use account balance. They have provider and per-user limits.

What is the API base URL?

The OpenAI-compatible base URL is https://api.unorouter.com/v1. API requests require a key created on the Tokens page.

How are paid models billed?

Paid requests deduct input and output token costs at the model's rates, or a flat per-call price. Failed requests are not billed.

Does UnoRouter support roleplay chat?

Yes. Its built-in chat includes characters, personas, lorebooks, cards, presets, group chats, scripting, memory, context and images.

Where is chat data stored?

The chat guide says everything is local, so users can take it anywhere and inspect anything.

Is UnoRouter open source?

The site says the entire stack is open source under OSI licenses and links to its GitHub organization.

UnoRouter

An open source AI API and a private chat app that share one key

AI Tools Developer Tools Open Source

Visit website

Launched: Nov 12, 2026
Website: unorouter.com
Pricing: Freemium
Platforms: Web, API, CLI, Android
Based in: Germany

UnoRouter is an open source AI platform for indie makers that pairs an OpenAI-compatible API with a chat app that keeps conversations in the browser, both running on the same account and key.

On the API side you set one base URL in your SDK, Claude Code, Cline or OpenCode and call models from many vendors by name. The gateway watches each upstream route and moves traffic to a healthy one when a provider errors or rate limits, so your project does not inherit somebody else's outage.

On the chat side you get a web client for talking to the same models, with history stored locally in your browser rather than on a server. Nothing to install, and nothing of yours sitting in a database you cannot see.

Pricing starts at zero: free models and a free tier, then pay as you go for paid models, with credit that you top up when you need it. All of the code, from the web app to the gateway, is public on GitHub.

What UnoRouter does

Multi-provider model gateway: Reach models from many providers through one OpenAI compatible API key.
Free and paid model access: Use models with a :free suffix or pay per token or per call for paid models.
Built-in character chat: Chat with characters using cards, personas, lorebooks, group chats, presets and branches.
Developer API and integrations: Connect coding agents, chat clients, roleplay apps and CLI tools through compatible endpoints.
Prompt scripting and automation: Use macros, regex scripts, triggers, Lua and JavaScript plugins around messages.
Usage controls and auditing: Manage key expiry, spending limits, model and IP allowlists, and inspect request logs.

Who it's for

Indie makers, developers using coding agents, and people who want a private web chat with many AI models

Why people use UnoRouter

Each AI vendor wants its own account, key, SDK and invoice, and a single vendor outage can take an AI feature down.

One OpenAI-compatible key for many vendors, automatic failover between upstream routes, and a browser based chat that runs on the same account.

The whole stack is open source, and the chat stores history locally in the browser, so how requests are routed and where your data lives can both be checked.

Pricing

UnoRouter is free to start, with paid plans from $1–$10000.

Pay as you go

$1–$10000

Add credit and use any model; credit does not expire.

Top-up options from $1 to $10,000
Use any model
Pay per token or per call
Purchased balance does not expire

Starter

$20/ month

Pay $20/month and receive $40 in API credit.

2x the API credit you pay for
$40 API credit per month
$10 delivered each week
50% less waiting on free models
Unused credit does not carry over weekly

Pro

$50/ month

Pay $50/month and receive $100 in API credit.

2x the API credit you pay for
$100 API credit per month
$25 delivered each week
75% less waiting on free models
Unused credit does not carry over weekly

Scale

$100/ month

Pay $100/month and receive $200 in API credit.

2x the API credit you pay for
$200 API credit per month
$50 delivered each week
No waiting between free model requests
Unused credit does not carry over weekly

Platforms

UnoRouter runs on Web, API, CLI, Android.

UnoRouter FAQ

What is UnoRouter?: An open source, OpenAI compatible AI gateway: one API key for models from many providers, plus a built in chat client.
Which models can be used?: The site lists models from OpenAI, Anthropic, Google, DeepSeek, Moonshot, Zhipu, Qwen and other providers.
Are there free models?: Yes. Free models use the :free suffix, route only to free providers and do not use account balance. They have provider and per-user limits.
What is the API base URL?: The OpenAI-compatible base URL is https://api.unorouter.com/v1. API requests require a key created on the Tokens page.
How are paid models billed?: Paid requests deduct input and output token costs at the model's rates, or a flat per-call price. Failed requests are not billed.
Does UnoRouter support roleplay chat?: Yes. Its built-in chat includes characters, personas, lorebooks, cards, presets, group chats, scripting, memory, context and images.
Where is chat data stored?: The chat guide says everything is local, so users can take it anywhere and inspect anything.
Is UnoRouter open source?: The site says the entire stack is open source under OSI licenses and links to its GitHub organization.