This project provides a reverse-engineered API for Kimi AI's long-context language model, enabling users to leverage its advanced capabilities like document summarization, web search, and image analysis. It's designed for developers and researchers who need programmatic access to Kimi AI's features without relying on the official platform, offering a free, zero-configuration deployment.

How It Works

The API acts as a proxy, intercepting requests and forwarding them to Kimi AI's backend using a user-provided refresh_token. It mimics the OpenAI API structure, specifically the /v1/chat/completions endpoint, allowing compatibility with existing OpenAI-compatible clients. The project supports various Kimi models (kimi, kimi-search, kimi-research, kimi-k1, kimi-math) and features like streaming output, multi-turn conversations, and agent integration.

Quick Start & Requirements

• Install/Run: Docker is the primary deployment method.

  docker run -it -d --init --name kimi-free-api -p 8000:8000 -e TZ=Asia/Shanghai vinlic/kimi-free-api:latest
  

• Prerequisites: A refresh_token obtained from the Kimi AI website (kimi.moonshot.cn).
• Setup: Minimal setup time, primarily involves obtaining the token and running the Docker container.
• Docs: Kimi AI Free API

Highlighted Details

• OpenAI API compatibility for seamless integration.
• Supports long-context understanding, web search, and image analysis.
• Offers multiple Kimi AI models and specialized modes (e.g., K1 thinking model).
• Enables multi-account access by concatenating refresh_tokens.

Maintenance & Community

The project is maintained by the LLM-Red-Team. Links to other similar reverse-engineered APIs for various LLMs are provided.

Licensing & Compatibility

The project is for research and learning purposes only. It explicitly prohibits commercial use or providing services to others, warning of potential risks and service pressure on the official platform.

Limitations & Caveats

This is a reverse-engineered API and is inherently unstable; users are strongly advised to use the official Moonshot AI platform for commercial use to avoid potential account bans. Token statistics are not available as inference is handled server-side.