plane

The Plane MCP Server exposes a programmable interface to manage Plane workspaces, projects, issues, labels, cycles, and more via the Model Context Protocol (MCP). This lets AI agents and automated tools interact with your Plane workspace to create projects, update work items, assign tasks, move issues through workflows, and query project metadata such as issue types, states, modules, and cycles. Use the provided tool list to call specific actions like creating a project, listing projects, retrieving user information, or updating issue states, all through a consistent MCP API surface. By integrating this server with your AI-powered assistants, automation scripts, or custom agents, you can drive Plane workflows in a structured, reproducible way while keeping your domain logic centralized in MCP calls.

Prerequisites:

• Node.js v14+ (recommended LTS) and npm installed on your system
• Access to the Plane MCP Server package registry (npm) or a locally available package

Option A: Install as a local npm package (recommended for development)

1. Install the package from npm: npm install plane-mcp-server --save
2. Navigate to your project directory where the server is installed and identify the entry file (commonly server.js or an equivalent main file).
3. Start the server: node server.js

Option B: Install globally for quick start (CLI usage)

1. Install globally: npm install -g plane-mcp-server
2. Run the server (CLI may expose a command like plane-mcp-server or similar; if not, use the local node runner): plane-mcp-server or, if a specific CLI is not provided, use: node $(npm -g prefix)/lib/node_modules/plane-mcp-server/server.js

Note: If the package exposes environment-based configuration, you may also configure via environment variables or a config file as documented by the package maintainers.

Environment variables you may encounter or want to set (placeholders):

• PLANE_API_TOKEN: auth token for Plane API access
• PLANE_ORG_ID: default organization ID to scope requests
• MCP_PORT: port on which the MCP server will listen
• MCP_LOG_LEVEL: logging level (info, debug, warn, error)
• MAX_CONCURRENT_REQUESTS: limit on parallel MCP calls Common issues:
• If the MCP server cannot connect to Plane, verify PLANE_API_TOKEN and network reachability to Plane API endpoints.
• Ensure the Node.js version matches the minimum requirement of the package and that any peer dependencies are resolved.
• When upgrading, review breaking changes in issue type/state/label parameter structures, as the API surface can be sensitive to exact field names.
• If running behind a reverse proxy, configure the MCP_PORT and related proxy settings accordingly.