ChongPlus Image MCP

An MCP server that lets AI assistants generate and edit images through the ChongPlus Image API.

Languages: English | 简体中文 | 繁體中文 | 日本語 | Español | Русский | 한국어

It is designed for normal chat usage: after installation, the assistant checks whether a key is configured, asks for one only when necessary, and saves it in the user's local configuration directory. Users do not need to set environment variables or run Python commands.

What the assistant can do

• Check whether ChongPlus is ready to use.

• Save or replace a ChongPlus API key securely.

• Generate one to four images from a prompt.

• Edit one local reference image with a prompt.

• Explain available image sizes and limits.

Generated images are saved locally. The tools return their absolute paths, so an MCP client can show or open them.

Related MCP server: openai-gpt-image-1-mcp

Install

The simplest route is to ask an MCP-capable AI client to install this repository. For manual configuration, use uvx:

{
  "mcpServers": {
    "chongplus-image": {
      "command": "uvx",
      "args": ["--from", "git+https://github.com/Rodert/chongplus-image-mcp", "chongplus-image-mcp"]
    }
  }
}

For a checked-out repository, replace the command and arguments with:

{
  "command": "uv",
  "args": ["run", "--directory", "/absolute/path/to/chongplus-image-mcp", "chongplus-image-mcp"]
}

First use

Ask the assistant to generate or edit an image. It should call setup_status first. If no key exists, it asks you for a ChongPlus API key and calls configure_api_key.

Get a key from https://api.chongplus.plus/keys; choose the image-generation group. Treat this as sensitive information: only provide it to an AI client and local MCP server you trust.

The key is stored at:

On macOS and Linux, the directory is set to 0700 and the file to 0600.

Tools

The server accepts existing local files for editing. It does not accept URLs, and it never returns file contents from arbitrary local paths.

Development

Requires Python 3.10 or newer and uv.

uv sync
uv run python -m unittest discover -s tests
uv run chongplus-image-mcp

Tests are offline and do not require an API key.

Errors

• 401 or 403: the API key may be invalid, lack image-model access, or have no remaining quota.

• 403 with Cloudflare 1010: ChongPlus's edge firewall blocked the request before the API processed it. Contact ChongPlus support with the approximate request time.

• Connection errors: check your internet connection and retry.

The API source of truth is https://api.chongplus.plus/tools/image-studio/docs/.