Deploy LiteLLM in 30 Seconds

LiteLLM is an open-source LLM gateway that puts one OpenAI-compatible API in front of 100+ providers, including OpenAI, Anthropic, Gemini, Bedrock, and Azure. It adds virtual keys, per-key budgets and rate limits, spend tracking, and an admin UI, and can route, load-balance, and fall back across providers. It is MIT-licensed with roughly 50,000 GitHub stars and is used by teams like Netflix and Stripe.

Yes. LiteLLM gives you the same single-endpoint, multi-provider experience as OpenRouter, but self-hosted on your own server. You bring your own provider API keys, your spend stays on your own Postgres, and there is no markup on tokens or third-party access to your requests. It also covers what tools like Helicone and Portkey do for spend tracking and key management, all from one $15/mo Grow plan.

Yes. LiteLLM is a gateway, not a model provider. You add your own keys for OpenAI, Anthropic, Gemini, Bedrock, Azure, and any other provider you want. LiteLLM stores them encrypted at rest, then routes, load-balances, and falls back across them behind one OpenAI-compatible endpoint.

Yes. Every pod is a private, single-tenant gateway with its own master key, its own Postgres database, and its own provider keys encrypted at rest. There are no public sign-ups, so only the keys you issue can reach the gateway.

LiteLLM logs every request and its token cost per virtual key in Postgres, and surfaces spend in the admin UI. You can set per-key budgets and rate limits, so a single key, team, or app cannot run past its cap. That gives you one place to see and govern spend across every provider you route through.

Self-hosting LiteLLM normally means provisioning a server, running Postgres, wiring up TLS, and managing the config. On InstaPods you skip all of that - click deploy, and LiteLLM is running with managed TLS and a backed-up Postgres in about 30 seconds. Then you add your provider keys and start issuing virtual keys.