Skip to content
Back to search

sourcerer-mcp

MCP for semantic code search & navigation that reduces token waste

Safety scan

A static text scan, not a guarantee. Select the badge to see what was scanned and every finding.

Checked today
Badge for your README
findsafeskills safety rating

Markdown

[![findsafeskills safety rating](/api/listings/st3v3nmw--sourcerer-mcp/badge.svg)](/listings/st3v3nmw--sourcerer-mcp)

HTML

<a href="/listings/st3v3nmw--sourcerer-mcp"><img src="/api/listings/st3v3nmw--sourcerer-mcp/badge.svg" alt="findsafeskills safety rating" /></a>

Install

i

Install instructions from the README

As written by the repo's author. findsafeskills has not verified these steps.

Go

go install github.com/st3v3nmw/sourcerer-mcp/cmd/sourcerer@latest

Homebrew

brew tap st3v3nmw/tap
brew install st3v3nmw/tap/sourcerer

What was scanned

The files the safety scan read, as written by the repository's author and shown as plain text. Lines with a finding are marked.

README.md · README · 110 lines

View this file on GitHub (read at the repository's current version: the scan did not record which commit it read)

  1. # Sourcerer MCP 🧙
  2. An MCP server for semantic code search & navigation that helps AI agents work
  3. efficiently without burning through costly tokens.
  4. Instead of reading entire files, agents can search conceptually and
  5. jump directly to the specific functions, classes, and code chunks they need.
  6. ## Demo
  7. [![asciicast](https://asciinema.org/a/736638.svg)](https://asciinema.org/a/736638)
  8. ## Requirements
  9. - **OpenAI API Key**: Required for generating embeddings (local embedding support planned)
  10. - **Git**: Must be a git repository (respects `.gitignore` files)
  11. - **Add `.sourcerer/` to `.gitignore`**: This directory stores the embedded vector database
  12. ## Installation
  13. ### Go
  14. ```shell
  15. go install github.com/st3v3nmw/sourcerer-mcp/cmd/sourcerer@latest
  16. ```
  17. ### Homebrew
  18. ```shell
  19. brew tap st3v3nmw/tap
  20. brew install st3v3nmw/tap/sourcerer
  21. ```
  22. ## Configuration
  23. ### Claude Code
  24. ```shell
  25. claude mcp add sourcerer -e OPENAI_API_KEY=your-openai-api-key -e SOURCERER_WORKSPACE_ROOT=$(pwd) -- sourcerer
  26. ```
  27. ### mcp.json
  28. ```json
  29. {
  30. "mcpServers": {
  31. "sourcerer": {
  32. "command": "sourcerer",
  33. "env": {
  34. "OPENAI_API_KEY": "your-openai-api-key",
  35. "SOURCERER_WORKSPACE_ROOT": "/path/to/your/project"
  36. }
  37. }
  38. }
  39. }
  40. ```
  41. ## How it Works
  42. Sourcerer 🧙 builds a semantic search index of your codebase:
  43. ### 1. Code Parsing & Chunking
  44. - Uses [Tree-sitter](https://tree-sitter.github.io/tree-sitter/) to parse source files into ASTs
  45. - Extracts meaningful chunks (functions, classes, methods, types) with stable IDs
  46. - Each chunk includes source code, location info, and contextual summaries
  47. - Chunk IDs follow the format: `file.ext::Type::method`
  48. ### 2. File System Integration
  49. - Watches for file changes using `fsnotify`
  50. - Respects `.gitignore` files via `git check-ignore`
  51. - Automatically re-indexes changed files
  52. - Stores metadata to track modification times
  53. ### 3. Vector Database
  54. - Uses [chromem-go](https://github.com/philippgille/chromem-go) for persistent vector storage in `.sourcerer/db/`
  55. - Generates embeddings via OpenAI's API for semantic similarity
  56. - Enables conceptual search rather than just text matching
  57. - Maintains chunks, their embeddings, and metadata
  58. ### 4. MCP Tools
  59. - `semantic_search`: Find relevant code using semantic search
  60. - `get_chunk_code`: Retrieve specific chunks by ID
  61. - `find_similar_chunks`: Find similar chunks
  62. - `index_workspace`: Manually trigger re-indexing
  63. - `get_index_status`: Check indexing progress
  64. This approach allows AI agents to find relevant code without reading entire files,
  65. dramatically reducing token usage and cognitive load.
  66. ## Supported Languages
  67. Language support requires writing [Tree-sitter queries](https://github.com/st3v3nmw/sourcerer-mcp/blob/main/internal/parser/go.go) to
  68. identify functions, classes, interfaces, and other code structures for each language.
  69. **Supported:** Go, JavaScript, Markdown, Python, TypeScript
  70. **Planned:** C, C++, Java, Ruby, Rust, and others
  71. ## Contributing
  72. All contributions welcome! See [CONTRIBUTING.md](CONTRIBUTING.md).
  73. ```
  74. $ ls @stephenmwangi.com
  75. - gh:st3v3nmw/obsidian-spaced-repetition
  76. - gh:st3v3nmw/lsfr
  77. ```

Ratings and comments

Sign in to rate this skill.

Sign in

Free to use. If it’s helped, pass it on.