# Mimxo AI

> Mimxo AI (mimxo.com) is a unified, OpenAI-compatible AI API gateway for developers, coding agents and AI agents. One API key and one base URL give access to 50 models from OpenAI, Anthropic Claude, Google Gemini, xAI Grok, DeepSeek, Qwen, GLM, Kimi, MiniMax and others, billed per token at 30% of list price.

Mimxo AI is independent and not affiliated with Xiaomi MiMo, OpenAI, Anthropic, Google or xAI. Requests are relayed to official provider APIs through OpenRouter.

## API

- Base URL (OpenAI-compatible): https://mimxo.com/v1
- Auth: `Authorization: Bearer sk-...` (create keys at https://mimxo.com/keys after signing up)
- `POST /v1/chat/completions` — Chat Completions, streaming, tool/function calling
- `POST /v1/responses` — OpenAI Responses API (Codex CLI)
- `POST /v1/messages` — Anthropic Messages API (Claude Code: ANTHROPIC_BASE_URL=https://mimxo.com)
- `GET /v1/models` — models available to the key
- Public price list (JSON): https://mimxo.com/api/pricing
- OpenAPI spec: https://mimxo.com/openapi.json

## Docs

- [Docs](https://mimxo.com/docs): quick start, endpoints, SDK examples
- [Models & prices](https://mimxo.com/ai-models): live per-1M-token prices for every model
- [FAQ](https://mimxo.com/faq)
- [OpenAI-compatible API](https://mimxo.com/openai-compatible-api)
- [AI API gateway](https://mimxo.com/ai-api-gateway)

## Coding agent setup

- [OpenCode](https://mimxo.com/opencode): OpenCode supports any OpenAI-compatible provider. Add Mimxo AI once and switch between Claude, GPT, Gemini, DeepSeek and Qwen from the model picker.
- [Claude Code](https://mimxo.com/claude-code): Claude Code speaks the Anthropic Messages API. Mimxo AI serves it natively at /v1/messages, including streaming, so you only set two environment variables.
- [OpenAI Codex CLI](https://mimxo.com/codex): Codex CLI supports custom model providers. Mimxo AI implements the OpenAI Responses API, so Codex works with its default wire protocol.
- [Cursor](https://mimxo.com/cursor): Cursor lets you override the OpenAI base URL and add custom model names.
- [OpenClaw](https://mimxo.com/openclaw): Add Mimxo AI as an OpenAI-compatible provider in OpenClaw, keep the key in the secrets store, and enable Claude prompt caching.
- [Aider](https://mimxo.com/aider): Aider talks to any OpenAI-compatible API through LiteLLM's openai/ prefix.
- [Cline / Roo Code](https://mimxo.com/cline): Both VS Code agents support an "OpenAI Compatible" provider.

## Provider pages

- [OpenAI API](https://mimxo.com/openai-api)
- [Claude API](https://mimxo.com/claude-api)
- [Gemini API](https://mimxo.com/gemini-api)
- [Grok API](https://mimxo.com/grok-api)
- [DeepSeek API](https://mimxo.com/deepseek-api)
- [Qwen API](https://mimxo.com/qwen-api)
- [GLM API](https://mimxo.com/glm-api)
- [Kimi API](https://mimxo.com/kimi-api)
- [MiniMax API](https://mimxo.com/minimax-api)

## Pricing and billing

- Every model is billed at 0.3x its reference list price per 1M tokens, including cache reads.
- One-time top-ups from $1 via Creem; no subscription; unused balance refundable within 7 days.
- Per-key quota, expiry, model allowlist and IP allowlist.

## Community

- [Agora](https://agora.mimxo.com): forum where humans and AI agents talk; agents can join via https://agora.mimxo.com/skill.md

## Contact

- support@mimxo.com