Jcodemunch MCP Server

GitHub source code retrieval using tree-sitter AST parsing to slash token usage by 95%.

Data last scanned today · Reviewed today

5350 tools★ 2,130PythonNOASSERTION

Overview

Jcodemunch MCP Server optimizes AI code exploration by utilizing tree-sitter AST parsing. By selectively retrieving specific code segments instead of entire files, it minimizes the volume of data sent to AI models, aiming to reduce context window usage significantly. It is compatible with a wide range of tools, including Claude Code, Cursor, VS Code, and various CLI interfaces. Developers can deploy it via pip or uvx, with options for standard PyPI installs or direct repository builds for version-stable environments.

Our verdict

This server carries a high-risk rating of 53/100. The score reflects significant security unknowns, specifically because its authentication method, OAuth support, and read-only implementation have not yet been editorially reviewed. Given the lack of a defined commercial license (NOASSERTION) and community-maintained status, users should exercise caution before integrating this into production environments handling sensitive intellectual property.

Reduces AI context token usage by up to 95 percent.
Zero direct dependencies minimize the software supply-chain surface area.
Active development with commits occurring daily.
Broad compatibility with popular AI coding assistants and CLI tools.

Security-critical features like authentication remain unreviewed.
License status is currently set to NOASSERTION.
Community-maintained rather than supported by an official vendor.

Tools

Tool	Description	Risk
find_references	Search for symbol references.	low
get_ranked_context	Query-driven token-budgeted context assembly.	low
find_dead_code	Identify unused code.	low
get_untested_symbols	Identify untested symbols.	low
get_changed_symbols	Map git diffs to specific code symbols.	low
get_symbol_importance	Rank symbols by architectural centrality using PageRank.	low
get_repo_map	Generate a signature-only repo overview ranked by PageRank.	low
find_similar_symbols	Detect consolidation candidates by finding similar code symbols.	low
get_group_contracts	Surface cross-repo API contracts.	low
find_implementations	Discover concrete implementations across multiple resolution sources.	low

Compatibility

Client	Local	Docker	Remote	Read-only
Claude Desktop
Cursor
VS Code
Windsurf
ChatGPT

Frequently asked questions

›What is the primary technical method used for code retrieval?

The server uses tree-sitter AST parsing to retrieve specific code elements such as functions, classes, methods, and constants with byte-level precision.

›What are the recommended ways to install the server?

You can install it using pip or uvx, or build from the repository source if you require version-stable deployments.

›Is jCodeMunch free to use?

The server is free for personal use, but commercial use requires a license.

›Does the server require specific dependencies for building from source?

Yes, installing directly from the repository requires git and builds the package from source.

›Which AI agents and IDEs are compatible with this server?

The server is compatible with any MCP-compliant client, including Claude Code, Cursor, VS Code, Windsurf, Continue, and the Codex CLI.