Add Sourcerer MCP

[st3v3nmw]

This PR adds the sourcerer MCP to the community servers list.

Description

Sourcerer is an MCP server for semantic code search & navigation that helps AI agents work efficiently without burning through costly tokens. Instead of reading entire files, agents can search conceptually and jump directly to the specific functions, classes, and code chunks they need.

Server Details

• Server: Sourcerer MCP
• Changes to: Adding a new server reference.

Motivation and Context

When AI agents work with codebases, they often end up reading entire files just to find a specific function or understand how something works. This burns through tokens quickly and can hit context limits on larger projects.

Sourcerer lets agents search for code semantically - they can ask for "authentication logic" or "database connection handling" and get back just the relevant chunks instead of whole files. It's been helpful for reducing token usage while still giving agents the precise code context they need.

How Has This Been Tested?

I've tested sourcerer extensively with Claude Code on Linux across various scenarios:

• Semantic search for finding specific functionality across large codebases
• Code navigation and chunk retrieval for targeted analysis
• Integration with different programming languages and project structures

The server also has comprehensive unit tests covering all supported languages (Python, JavaScript, TypeScript, etc.) to ensure reliable parsing and indexing across different codebases.

Breaking Changes

None. This is a new server reference that does not affect existing servers.

Types of changes

• Bug fix (non-breaking change which fixes an issue)
• New feature (non-breaking change which adds functionality)
• Breaking change (fix or feature that would cause existing functionality to change)
• Documentation update

Checklist

• I have read the MCP Protocol Documentation
• My changes follows MCP security best practices
• I have updated the server's README accordingly
• I have tested this with an LLM client
• My code follows the repository's style guidelines
• New and existing tests pass locally
• I have added appropriate error handling
• I have documented all environment variables and configuration options

Additional context

Sourcerer is written in Go and uses tree-sitter for reliable parsing across multiple languages. It maintains stable chunk IDs that persist across minor code changes and provides surgical precision for code retrieval.

Installation is available via go install or Homebrew. The server currently requires an OpenAI API key for embeddings (local embedding support is planned) and works within git repositories, automatically respecting .gitignore files. Configuration is straightforward with Claude Code using the claude mcp add command or standard mcp.json setup.