# warpllm

> A warp-speed, robust AI gateway for Rust, Node, and Python applications.

- [warpllm](https://warpllm.mintlify.app/index.md): One request shape, many providers. A Rust core compiled for Rust, Node, and Python, plus a self-hosted OpenAI-compatible gateway.
- [Quickstart](https://warpllm.mintlify.app/quickstart.md): Install warpllm, export one API key, and make your first routed request.
- [Installation](https://warpllm.mintlify.app/get-started/installation.md): Packages, supported runtimes, and prebuilt platform binaries.
- [Authentication](https://warpllm.mintlify.app/get-started/authentication.md): How warpllm finds provider API keys, and which variable each provider reads.
- [Model routing](https://warpllm.mintlify.app/concepts/model-routing.md): How provider/model strings resolve, what the registry guarantees, and why unknown names are errors.
- [Requests and responses](https://warpllm.mintlify.app/concepts/requests-and-responses.md): The chat-completion request shape, what warpllm models, and what it forwards untouched.
- [Errors](https://warpllm.mintlify.app/concepts/errors.md): One failure taxonomy across every provider, spoken in OpenAI's vocabulary.
- [Timeouts](https://warpllm.mintlify.app/concepts/timeouts.md): The total deadline, the stream read timeout, and which one you actually want.
- [Streaming](https://warpllm.mintlify.app/guides/streaming.md): Consume a reply as it is generated, and tell a complete stream from a truncated one.
- [Tool calling](https://warpllm.mintlify.app/guides/tool-calling.md): Let a model call your functions, and feed the results back into the conversation.
- [Structured outputs](https://warpllm.mintlify.app/guides/structured-outputs.md): Ask for JSON, or for JSON matching a schema you supply.
- [Migrating from the OpenAI SDK](https://warpllm.mintlify.app/guides/migrate-from-openai.md): Two routes off the official SDK: swap the base URL, or swap the client.
- [SDK reference](https://warpllm.mintlify.app/sdk/reference.md): The full client surface in Python, TypeScript, and Rust, on one page.
- [Model roster](https://warpllm.mintlify.app/models/overview.md): Every model warpllm can route to, with its published limits.
- [Add a model or provider](https://warpllm.mintlify.app/models/add-a-model.md): Edit one YAML file, cite the provider's docs, and let the tests check your work.
- [Gateway overview](https://warpllm.mintlify.app/gateway/overview.md): An OpenAI-compatible HTTP server in front of every provider warpllm routes to.
- [Run the gateway](https://warpllm.mintlify.app/gateway/run.md): Build and start warpllm-server with Docker or cargo, then make a request.
- [Gateway configuration](https://warpllm.mintlify.app/gateway/configuration.md): Every flag warpllm-server accepts, and the environment it reads.
- [Connecting a client](https://warpllm.mintlify.app/gateway/openai-clients.md): Point the official OpenAI SDKs, curl, or anything OpenAI-compatible at a running gateway.
- [HTTP API](https://warpllm.mintlify.app/api-reference/introduction.md): The OpenAI-compatible surface of a self-hosted warpllm gateway.
- [Create a chat completion](https://warpllm.mintlify.app/api-reference/create-chat-completion.md): Route one chat completion to the provider named by the model string.
- [Liveness check](https://warpllm.mintlify.app/api-reference/health.md): Report that the gateway process is serving, and which warpllm version it was built from.

## OpenAPI Specs

- [openapi](/api-reference/openapi.json)

## Optional

- [GitHub](https://github.com/warpllm/warpllm)
- [Changelog](https://github.com/warpllm/warpllm/blob/main/CHANGELOG.md)
- [Discord](https://discord.gg/payU7Vzuq5)
- [Reddit](https://www.reddit.com/r/warpllm/)


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.