mimocode2api

MiMoCode2API is a small API gateway written in Go that translates requests from the OpenAI chat format to MiMo's AI model service and back. The README is written in Chinese.

The practical use case is compatibility: many AI coding tools and applications (such as Cursor, Continue, and Hermes Agent) are built to speak to OpenAI's API format. MiMo is a separate AI language model with its own API. This gateway sits in between, accepting requests in the OpenAI format and forwarding them to MiMo, so you can point existing OpenAI-compatible tools at this gateway without modifying them. It supports both streaming responses (where text appears word-by-word as the model generates it) and standard non-streaming responses (where you wait for the full reply).

Deployment is done with Docker. The compiled Docker image is about 12 MB. When you start it for the first time, it automatically generates an API key that you use to authenticate requests to the gateway. The gateway listens on port 10000 by default. Configuration is done through environment variables, including the upstream MiMo API address, the port to listen on, and a debug logging toggle.

The supported model listed is mimo/mimo-auto, which accepts text and image input, supports up to one million tokens of context, and can output up to 128,000 tokens. The README includes a disclaimer (in Chinese) stating the project is for learning and research only, not commercial use, and that users bear responsibility for any consequences of using it. Licensed under MIT.