# DodoRouter

> DodoRouter is an OpenAI- and Anthropic-compatible LLM proxy with automatic provider fallback. Point any client at it once; it retries the next configured provider automatically on rate limits, outages, or context overflows, converts the response back to your client's format, and logs every request.

This file follows the [llms.txt](https://llmstxt.org) convention: a condensed, link-based index for AI agents and LLMs. For the complete documentation concatenated as plain Markdown in one file (no need to follow links), fetch [https://dodorouter.com/llms-full.txt](https://dodorouter.com/llms-full.txt) instead. Every page linked below also has its own raw-Markdown twin at the same URL with `.md` appended (e.g. `https://dodorouter.com/docs/quickstart.md`).


## Get Started

- [Overview](https://dodorouter.com/docs/): What DodoRouter is, how a request flows through it, and where to go next.
- [Quickstart (hosted)](https://dodorouter.com/docs/quickstart/): Register, create a router, add a provider key, and send your first request against the hosted service.
- [Self-hosting](https://dodorouter.com/docs/self-hosting/): Prerequisites, environment variables, and first boot for running your own DodoRouter instance.


## Core Concepts

- [Core Concepts](https://dodorouter.com/docs/concepts/): Routers, routing steps and fallback, provider keys vs router keys, plan types, reasoning effort, sessions, recordings, and logs.


## API Reference

- [API Reference](https://dodorouter.com/docs/api/): Base URL, authentication, the full endpoint list, why the model field is ignored, and the error format.
- [Chat Completions (OpenAI format)](https://dodorouter.com/docs/api/chat-completions/): The OpenAI-compatible chat/completions endpoint — request, verified response, and streaming.
- [Messages (Anthropic format)](https://dodorouter.com/docs/api/messages/): The Anthropic-compatible messages endpoint — request, verified response, and how it powers Claude Code with non-Anthropic models.
- [Responses API](https://dodorouter.com/docs/api/responses/): The Responses API endpoint used by Codex-style clients — request, streaming event sequence, and verified response.


## Providers & Models

- [Supported Providers](https://dodorouter.com/docs/providers/): Every LLM provider adapter DodoRouter supports, with auth type and provider-specific notes.
- [Models & Pricing](https://dodorouter.com/docs/models/): How DodoRouter syncs model capabilities and pricing, and how subscription keys are costed.


## Dashboard

- [Dashboard Guide](https://dodorouter.com/docs/dashboard/): A tour of every page in the DodoRouter web dashboard.


## Connect a Coding Agent

- [Connect a Coding Agent](https://dodorouter.com/docs/integrations/): Point Claude Code, OpenCode, Forge, or Codex CLI at DodoRouter — verified configs for each.
- [Claude Code](https://dodorouter.com/docs/integrations/claude-code/): Point Claude Code at DodoRouter via ANTHROPIC_BASE_URL and ANTHROPIC_AUTH_TOKEN.
- [OpenCode](https://dodorouter.com/docs/integrations/opencode/): Add DodoRouter as a custom OpenAI-compatible provider in opencode.jsonc.
- [Forge](https://dodorouter.com/docs/integrations/forge/): Configure Forge's built-in openai_compatible provider to point at DodoRouter.
- [Codex CLI](https://dodorouter.com/docs/integrations/codex-cli/): Add DodoRouter as a custom model_providers entry in ~/.codex/config.toml.


## Operations

- [Deployment](https://dodorouter.com/docs/deployment/): Elixir release boot sequence, migrations, and health checks for production.
- [Troubleshooting & FAQ](https://dodorouter.com/docs/troubleshooting/): Common errors and questions, and what actually causes each one.


## Optional

- [Full documentation, one file](https://dodorouter.com/llms-full.txt): every page above, concatenated as plain Markdown.
- [Source code](https://github.com/foxwise-ai/dodorouter): the Elixir/Phoenix implementation — the ground truth when this file and the code disagree.
- [Hosted service](https://dodorouter.com): register, subscribe, and manage routers without self-hosting.