home / mcp / pathfinder mcp server
Provides an MCP server workflow for research, planning, and implementation with session management and templates.
Configuration
View docs{
"mcpServers": {
"jamesctucker-pathfinder-mcp": {
"command": "uvx",
"args": [
"pathfinder-mcp"
],
"env": {
"PATHFINDER_HOST": "127.0.0.1",
"PATHFINDER_PORT": "8080",
"PATHFINDER_TRANSPORT": "stdio",
"PATHFINDER_SESSIONS_DIR": "/Users/yourusername/.pathfinder-sessions"
}
}
}
}Pathfinder MCP Server provides a three‑phase workflow—Research, Plan, and Implement—with explicit human approvals at each transition. It helps you focus on complex coding tasks by organizing information gathering, structured planning, and progress‑tracked execution in a repeatable, auditable way.
You connect with Pathfinder MCP Server through an MCP client to drive a session from research through implementation. Start a session to begin gathering research findings, then plan your implementation with an explicit plan, and finally execute the plan while tracking progress. You can compact the session context to keep long runs lightweight and reuse its summaries for future work.
Prerequisites: you need Node.js installed to run the MCP server via the command line tool, and you may use a local development flow if you plan to modify the server.
Install and run Pathfinder MCP Server in two approaches: a published, direct run and a local development setup.
Pathfinder MCP Server exposes its behavior through environment variables and MCP configurations. Use the environment variables to tailor transport, host, ports, and session storage. You can provide multiple MCP configurations to connect via different runtimes or transport methods.
Each session stores its research notes, planning templates, progress, and a compacted summary in a dedicated directory. This enables you to resume work, audit decisions, and share a clear history of how your implementation evolved.
If you are developing Pathfinder MCP Server locally, you can clone the project, install dependencies, and run the local server for testing and iteration. Use the provided development commands to run tests, lint, and build as you refine the code.
Server health and context status
Initialize research phase
Save research findings
Transition to planning and create plan template (requires approval)
Save the implementation plan
Execute implementation phase and progress tracking
Compress session context when utilization is high