Code Pathfinder
Index your Python or Go repo locally so your agent can trace call graphs, infer types, and search symbols instead of grepping blind.
Overview
Code Pathfinder is a MCP server for the Build phase that provides call graphs, type inference, and symbol search for local Python and Go projects.
What is this MCP server?
- Call graphs, type inference, and symbol search for Python and Go
- PyPI package codepathfinder v2.1.1 with `serve` and optional `--project` root
- Docker image docker.io/shivasurya/code-pathfinder:v2.1.1 as OCI alternative
- stdio MCP—indexes project root (defaults to current directory)
- Version 2.1.1 on PyPI and OCI docker.io/shivasurya/code-pathfinder:v2.1.1
- Languages supported: Python and Go per description
- Three capability areas: call graphs, type inference, symbol search
Community signal: 137 GitHub stars.
What problem does it solve?
Agents waste tokens and miss call sites when they cannot see structured graphs and types across Python or Go codebases.
Who is it for?
Solo developers on Python/Go repos who want local, stdio MCP code intelligence without sending source to a hosted analyzer.
Skip if: Rust/TypeScript-only shops, or teams unwilling to run a local indexer on each machine.
What do I get? / Deliverables
After you run `serve` against your project root, the agent can query symbols and call relationships grounded in your indexed repo.
- Project-indexed MCP tools for graphs, types, and symbols
- Faster agent navigation of Python/Go code during implementation
Recommended MCP Servers
Journey fit
Local code intelligence MCP is adopted during Build when you equip the agent to navigate a real codebase safely. Agent-tooling is the shelf for stdio MCP servers that run beside the repo and power navigation during implementation.
How it compares
Local stdio code-intelligence MCP, not a multi-app OAuth hub or hosted Rust crate oracle.
Common Questions / FAQ
Who is dev.codepathfinder/pathfinder for?
Builders using MCP agents on Python or Go codebases who need call graphs, types, and symbol search from a local indexer.
When should I use dev.codepathfinder/pathfinder?
Use it during Build agent-tooling setup when refactoring, onboarding the agent to a monorepo, or tracing impact before larger backend changes.
How do I add dev.codepathfinder/pathfinder to my agent?
Install codepathfinder 2.1.1 from PyPI or run the v2.1.1 Docker image, configure stdio MCP with command `serve` and `--project` set to your repo root.