Overview
What is Mantis MCP Server?
Mantis MCP Server is a service based on the Model Context Protocol (MCP) that integrates with the Mantis Bug Tracker system. It provides tools to query and analyze data from a Mantis instance via the MCP protocol.
How to use Mantis MCP Server?
Install the package with npm (globally or locally), configure environment variables (MANTIS_API_URL, MANTIS_API_KEY) in a .env file, and then add the server configuration to your MCP client (e.g., Cursor, VS Code) using the provided JSON examples for Windows or macOS/Linux.
Key features of Mantis MCP Server
- Issue management with filtering and detail retrieval
- User lookup by name and full user list
- Project list retrieval
- Issue statistics grouped by status, priority, severity, handler, or reporter
- Assignment statistics with status filtering
- Performance optimizations: field selection, pagination, auto-compression
Use cases of Mantis MCP Server
- Query and filter Mantis issues from an AI assistant
- Get user and project information without manual API calls
- Generate issue statistics for reporting and analysis
- Analyze issue assignment distribution across users
FAQ from Mantis MCP Server
What do I need to configure?
You must set MANTIS_API_URL (your Mantis API endpoint) and MANTIS_API_KEY (API token obtained from your MantisBT account).
How do I get a Mantis API key?
Log in to your MantisBT account, go to "My Account" > "API Tokens" tab, create a new token, and copy it.
Can I use the server without global installation?
Yes, on macOS/Linux you can use npx to run the latest version without global install; on Windows global installation is shown.
Where are logs stored if file logging is enabled?
Logs are saved to logs/mantis-mcp-server-combined.log (all levels) and logs/mantis-mcp-server-error.log (errors only), with a 5 MB file size limit and 5 historical files retained.
What transports does the server support?
The server uses standard input/output (stdio) as the transport for MCP, as shown in the configuration examples for Cursor and VS Code.
