MCP (Model Context Protocol) server for aicommit - AI-powered git commit message generation.
This MCP server allows AI assistants (like Claude) to generate git commit messages for your code changes. It acts as a bridge between LLMs and the aicommit tool, enabling AI-powered version control workflows.
There are two versions of this package available:
1. Regular Version @suenot/aicommit-mcp
The standard package that requires aicommit to be installed separately.
π¦ NPM Package: @suenot/aicommit-mcp
2. Bundled Version @suenot/aicommit-mcp-bundled
The bundled package that automatically installs aicommit for you.
π¦ NPM Package: @suenot/aicommit-mcp-bundled
| Feature | Regular | Bundled |
|---|---|---|
| Generate AI commit messages | β | β |
| Execute aicommit operations | β | β |
| Check git repository status | β | β |
| View aicommit providers | β | β |
| Automatic staging, pushing, pulling | β | β |
| Automatic installation of aicommit | β | β |
# Install aicommit first
npm install -g @suenot/aicommit
# or
cargo install aicommit
# Install the MCP server
npm install -g @suenot/aicommit-mcp# Install the bundled MCP server
npm install -g @suenot/aicommit-mcp-bundledYou can also run the MCP server using Docker:
# Pull the Docker image
docker pull suenot/aicommit-mcp
# Run the container
docker run -p 8888:8888 -v $(pwd):/workspace --name aicommit-mcp suenot/aicommit-mcp# Pull the Docker image
docker pull suenot/aicommit-mcp-bundled
# Run the container
docker run -p 8888:8888 -v $(pwd):/workspace --name aicommit-mcp-bundled suenot/aicommit-mcp-bundledThis MCP server is compatible with Smithery and can be deployed to the Smithery marketplace. The server implements "lazy loading" of the aicommit tool, which means:
- Tool listing doesn't require authentication or configuration
- Authentication and dependencies are only checked when tools are actually called
- The server supports deployment and tool display in the Smithery web interface
# Using the Smithery CLI
npx -y @smithery/cli@latest install @suenot/aicommit-mcp --client claude --config '{}'
# For bundled version
npx -y @smithery/cli@latest install @suenot/aicommit-mcp-bundled --client claude --config '{}'Both versions of the aicommit MCP server support direct installation from the Smithery web interface:
- Visit Smithery
- Click "Continue Anyway" when prompted
- Select your preferred client (Claude, Cursor, etc.)
- Click "Install Server"
The installer will guide you through the process:
- For the regular version, you'll need to install aicommit separately
- For the bundled version, aicommit will be installed automatically
If you encounter the "Uh oh! This server works best locally..." error:
- Ensure you have Node.js 14+ installed
- Install the package manually using npm:
npm install -g @suenot/aicommit-mcp # or for bundled version npm install -g @suenot/aicommit-mcp-bundled - Configure your MCP client (Claude, Cursor, etc.) as shown in the Assistant Integration section
{
"mcpServers": {
"aicommit": {
"command": "npx",
"args": ["-y", "@suenot/aicommit-mcp"]
// For bundled version: "args": ["-y", "@suenot/aicommit-mcp-bundled"]
}
}
}To use Docker with Claude Desktop, add this to your configuration:
{
"mcpServers": {
"aicommit": {
"command": "docker",
"args": ["run", "--rm", "-v", "${workspaceFolder}:/workspace", "suenot/aicommit-mcp"]
// For bundled version: "args": ["run", "--rm", "-v", "${workspaceFolder}:/workspace", "suenot/aicommit-mcp-bundled"]
}
}
}{
"mcpServers": {
"aicommit": {
"command": "npx",
"args": ["-y", "@suenot/aicommit-mcp"]
}
}
}{
"mcpServers": {
"aicommit": {
"command": "npx",
"args": ["-y", "@suenot/aicommit-mcp"]
}
}
}{
"mcpServers": {
"aicommit": {
"command": "npx",
"args": ["-y", "@suenot/aicommit-mcp"]
}
}
}Install via Smithery CLI:
# Regular version
npx -y @smithery/cli@latest install @suenot/aicommit-mcp --client claude --config '{}'
# Bundled version
npx -y @smithery/cli@latest install @suenot/aicommit-mcp-bundled --client claude --config '{}'For complete documentation for each package, please refer to:
MIT
- aicommit - The CLI tool this MCP server uses
- Model Context Protocol - Learn more about MCP