probe
Semantic code search for AI — ripgrep speed with tree-sitter AST parsing.
jcodemunch-mcp est un projet de développement open source, créé par jgravelle (jgravelle/jcodemunch-mcp sur GitHub). Il compte 2 740 étoiles sur GitHub et 372 forks, il est publié sous la licence NOASSERTION et a été mis à jour pour la dernière fois sur GitHub le 8 octobre 2026.
Symbol-level code retrieval MCP server (tree-sitter AST) built to slash agent token costs on exploration
Point it at a repo your agent explores often, where whole-file reading is burning context and cost.
Install with uv tool, or run zero-install via uvx.
Run jcodemunch-mcp init to auto-detect and configure your MCP clients and prompt policy, or add it manually (e.g. claude mcp add).
Ask your agent to index the repo, then find a function by description and fetch only its source.
get_symbol_source returns the exact function body, byte-precise, instead of the whole file.
find_importers, get_blast_radius, get_call_hierarchy, find_dead_code, get_changed_symbols, get_hotspots and search_ast answer questions grep cannot.
assemble_task_context and plan_turn classify intent and gather the right symbols under one token budget.
check_edit_safe, check_delete_safe, check_rename_safe and get_pr_risk_profile return explicit verdicts, including null when the index cannot support a claim.
Calibrated confidence scores, freshness flags, coverage contracts on absence claims, and automatic secret redaction before content reaches the LLM.
Tested configs for many clients (Claude Code, Cursor, Codex, Cline, Windsurf, Gemini CLI and more) and 70+ tree-sitter languages, with monorepo support.
Clone jCodeMunch if you want the source; the normal install is step 2.
git clone https://github.com/jgravelle/jcodemunch-mcp.git
cd jcodemunch-mcpInstall the server and let init configure your detected MCP clients.
uv tool install jcodemunch-mcp
jcodemunch-mcp initThe zero-install alternative from the README uses uvx.
claude mcp add -s user jcodemunch -- uvx jcodemunch-mcpjgravelle maintains it; the product site is jcodemunch.com.
The README says it is free for personal use, with commercial licenses for using it to make money. Check the license terms before commercial use — GitHub lists the license as NOASSERTION.
The server itself is a local retrieval tool; your agent's model costs are separate. Some optional extras use AI summaries per provider, per the README.
Both give agents AST-based code search. jCodeMunch focuses on symbol-exact retrieval and structural/safety tooling for MCP clients; ccc is a CLI-first engine with Skill/MCP modes and local embeddings.
uv tool install jcodemunch-mcp, run jcodemunch-mcp init, then ask your agent to index the repo and fetch a symbol.
Semantic code search for AI — ripgrep speed with tree-sitter AST parsing.
A 25MB database client for 100+ databases — with a built-in AI assistant, MCP server and CLI
Code-search MCP for Claude Code: make your entire codebase the context for any coding agent.
Kimi Code CLI — Moonshot's terminal coding agent: single-binary TUI with video input, subagents and MCP.
Generates agent-first Go CLIs (plus MCP servers) for any API or website — with local SQLite sync built in
An MCP server plus a CLI tool that indexes local code into a graph database to provide context to AI assistants.
Les autres projets de la catégorie Développement avec le plus d'étoiles du catalogue LibreAI — la sélection réaliste d'alternatives si jcodemunch-mcp ne vous convient pas.
The open source coding agent
The model-definition framework behind modern AI — thousands of pretrained transformers for text, vision, and audio.
Anthropic's agentic coding tool that lives in your terminal and works across your whole codebase
Cross-platform desktop all-in-one manager for Claude Code, Codex, OpenCode and other coding agents
OpenAI's lightweight coding agent that runs locally in your terminal
Projets de la catégorie Développement proches dans le catalogue, les plus proches en étoiles de jcodemunch-mcp — continuez à explorer de part et d'autre.
Lightweight AST-based semantic code search CLI (ccc) that makes coding agents faster and cheaper
Claude Code skill: four-phase plan hardening (recon, interrogate, Codex adversarial review, cross-model build & inspection) — two AI models harden your plan …
A bridge between Streamable HTTP / SSE and stdio MCP transports — connect any client to any server
A coding-agent harness optimized for smaller local LLMs — small models, serious coding work
Learn to code securely while having fun through our popular open source in-editor experience, designed for developers, students, and anyone curious about sec…
Research into how agentic AI coding assistants work. Reconstructed prompt patterns, agent coordination, and security classification
Gratuit et ouvert : récupérez le code source, mettez une étoile au dépôt et commencez à créer.