Skip to main content
Glama

repo-map-mcp

MCP server that maintains a self-updating, per-file JSON index of a repo (directory, filename, description, function names, timestamps) so an agent can query "what's probably relevant" instead of grepping/reading the whole tree.

File exploration and path annotation computed locally; Description and file purpose are dictated by the agent, which provides them even better context.

Install

uv sync

Related MCP server: Moxie Docs MCP

Tools

  • index_repository(root_path, force=False) — walk root_path (respecting .gitignore + hardcoded excludes), hash each file, re-parse function names for changed files, update the on-disk index at <root_path>/.repo-index/index.json. Returns a summary string.

  • get_repo_map(root_path, dir_prefix=None) — return the JSON index, optionally scoped to a subdirectory.

  • search_files(root_path, query) — substring match against filename, description, and function names.

  • get_stale_files(root_path) — files whose descriptionStale flag is set, with their function names, so the calling agent can write a description without opening the file.

  • set_description(root_path, path, description) — writes the description back and clears the stale flag.

Supported languages for function extraction (via tree-sitter): Python, JavaScript, TypeScript, TSX, Java, Rust. Other file types still get indexed, with funcs: [].

Requirements:

  • Python

  • uv

Run

uv run repo-map-mcp

Runs on stdio — add it to an MCP client config (e.g. Claude Code) pointing command at uv run repo-map-mcp (cwd set to this directory, or use an absolute path).

Test

uv run pytest -q
uv run python3 scripts/smoke_client.py   # end-to-end tool call smoke test

Available Tools

5 tools
get_repo_mapA

Return the JSON repo index for root_path, optionally scoped to dir_prefix. Call this before exploring.

Inputs: root_path (required), dir_prefix. Returns: result

get_stale_filesB

Return files in root_path's index whose description is stale (funcs changed), with their funcs.

Inputs: root_path (required). Returns: result

index_repositoryB

Walk root_path, hash/parse changed files, update the .repo-index/index.json map. Returns a summary.

Inputs: root_path (required), force. Returns: result

search_filesA

Substring search over fileName/description/funcs in the repo index for root_path.

Inputs: query (required), root_path (required). Returns: result

set_descriptionB

Write a one-line description for path in root_path's index and clear its stale flag.

Inputs: path (required), root_path (required), description (required). Returns: result

Tool Schema Changelog

Recent tool additions, removals, and schema changes observed during successful MCP inspections.

Loading tool schema changelog…

    TDQS

    A3.8/5.0

    Scored across 5 tools

    Disambiguation5/5

    Each tool has a distinct responsibility: setting descriptions, indexing, fetching the map, searching, and retrieving stale files. No overlap or ambiguity exists between them.

    Naming Consistency5/5

    All tools follow a consistent snake_case verb_noun pattern (set_description, index_repository, get_repo_map, search_files, get_stale_files). Predictable and uniform.

    Tool Count5/5

    5 tools is a well-scoped size for a focused repository indexing server. Each tool earns its place with a clear role, no redundancy.

    Completeness5/5

    The tool surface covers the core workflow: indexing, description management, retrieval, search, and staleness detection. No obvious gaps; the lifecycle is fully addressed.

    Maintenance

    ActivityInactive
    ResponsivenessNo recent community issues

    Related MCP Connectors

    Related MCP Servers

    • A
      license
      Not graded
      quality
      D
      maintenance
      Enables AI agents to intelligently navigate and understand codebases by providing instant file descriptions, semantic search, and context-aware recommendations, eliminating the need to repeatedly scan files.
      135 PyPI
      23
      MIT
    • A
      license
      Not graded
      quality
      D
      maintenance
      Enables LLM agents to efficiently understand and navigate a codebase by providing semantic search over symbols and a reference graph, replacing expensive grep/glob calls with structured tools like definition lookup, caller/callee queries, and change-impact analysis.
      4
      MIT
    • A
      license
      Not graded
      quality
      C
      maintenance
      Enables agentic coding tools to maintain a living, in-repo model of intent, architecture, and code, with deterministic indexing, bidirectional linking, and continuous drift detection for task-scoped context and reconciliation.
      MIT