Self-hosting

Yolorouter is an Apache-2.0 licensed LLM gateway that runs entirely on your own machine. The admin console is embedded in the binary: no Node runtime, no separate frontend deploy. SQLite works out of the box, PostgreSQL when you want it.

What you get

• Four wire protocols, freely interconverted — OpenAI Chat Completions, OpenAI Responses, Anthropic Messages, and Gemini generateContent. When the caller's protocol differs from the provider's, the request is translated on the way out — streaming event syntax included.
• Multi-provider failover — map one public model name to an ordered list of provider candidates. When one is down, requests fail over to the next, and the caller never sees a different model name.
• Upstream key rotation — give each provider a pool of keys. Rate-limited, unauthorized, or quota-exhausted keys are skipped automatically.
• Per-key access control — model allowlists, rate and concurrency limits, cumulative budget caps, optional expiry, and instant revocation.
• Built-in observability — token and cost KPIs, usage broken down by model, provider and time, and request logs with the full per-attempt routing chain. Any view exports to CSV.
• Image & video generation — the OpenAI Images and Videos APIs beside chat, billed per delivered image (quality×size table) or per delivered second (resolution-tier table). Non-OpenAI-shaped media providers are spoken natively — DashScope, Volcengine Ark, Kling, and MiniMax — and keys on media-only accounts verify through real media probes. See Image generation and Video generation.
• Cost optimization — custom system prompt injection and compression of bulky tool output, with the console reporting what each actually saved.

How this relates to YoloRouter Cloud

Self-hosting means bringing your own upstream API keys. If you would rather not sign up and top up with every provider separately, YoloRouter Cloud can serve as one of those upstreams — it ships in the admin console's provider preset list, so you pick it and paste a key.

The two are not mutually exclusive: a self-hosted instance can mix direct provider accounts with YoloRouter Cloud, and your failover rules decide the order.

Next steps

• Installation — Docker, the one-line installer, release binaries, building from source
• Configuration — every config.yaml field
• Updating and uninstalling
• Architecture

Source and issues live on GitHub.