killop

killop / codedb-mcp

Public

fast code databse mcp

13
1
89% credibility
Found May 22, 2026 at 13 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
Rust
AI Summary

codebase-mcp is a local code search and analysis tool that works with AI assistants. It indexes your source code, builds a map of how your files and symbols connect, and provides fast search capabilities. You install a skill package, it sets up locally in your project, and your AI assistant can then answer questions about your code by looking things up in the index rather than re-reading everything. The tool supports multiple programming languages and keeps everything organized in a project-local folder.

How It Works

1
🔍 You have a big codebase

Your project has grown so large that AI assistants struggle to understand it or find things quickly.

2
📦 You install a skill

You copy a ready-made skill package into your AI assistant that teaches it to work with your specific code.

3
🗺️ Your code gets mapped

The skill scans every file, understands the structure, and builds a searchable map of your entire project including how files depend on each other.

4
🔄 Everything stays in sync

As you edit files, the map updates automatically so your assistant always has the latest picture of your project.

5
What do you want to find?
🔤
Search for code

Find any text, pattern, or concept across thousands of files instantly

🔗
Trace connections

See which files use a function, or what a file depends on

📊
Explore structure

Understand how your code is organized into natural groups and modules

⚡ You get instant answers

Instead of searching blindly, your assistant navigates your code map and gives you precise answers with full context in seconds.

Sign up to see the full architecture

4 more

Sign Up Free

Star Growth

See how this repo grew from 13 to 13 stars Sign Up Free
Repurpose This Repo

Repurpose is a Pro feature

Generate ready-to-use prompts for X threads, LinkedIn posts, blog posts, YouTube scripts, and more -- with full repo context baked in.

Unlock Repurpose
AI-Generated Review

What is codedb-mcp?

codedb-mcp is a Rust-based MCP server that turns your codebase into a searchable, graph-aware knowledge base. Instead of grepping through files manually, you get semantic search, symbol outlines, caller tracking, dependency graphs, and batch operations through a standardized tool interface. It parses your source with tree-sitter, builds BM25 and vector indexes, constructs a code graph with community detection, and serves everything through the Model Context Protocol. The result: sub-10ms search times on codebases with tens of thousands of files, beating ripgrep in most benchmarks while adding semantic understanding.

Why is it gaining traction?

The hook is hybrid search that actually works on code. BM25 plus vector embeddings means you can search for "network listener manager" and get relevant results even without exact matches. The dependency graph with Louvain communities surfaces architectural boundaries you might not have consciously noticed. Batch operations let you run up to 100 tool calls in a single MCP round trip, which matters when you're building AI agents that need fast, code-aware tooling. The skill-first distribution model is also clever: copy a directory, run a setup script, and your agent gets search superpowers without global installs.

Who should use this?

AI agent developers building code-understanding workflows will get the most value. If you're using Codex, Claude Code, or similar tools and want them to navigate large codebases faster, this delivers. Backend developers maintaining C# or Java monoliths will appreciate the typed caller tracking and namespace-aware dependency analysis. Teams evaluating "fast code" alternatives for code review or refactoring tooling will find the benchmarks compelling. It's less useful for small projects or developers who just need occasional grep.

Verdict

This is a technically impressive project with solid benchmarks and a clear use case, but the credibility score of 0.8999% and 13 stars signal it's early-stage and community validation is minimal. The documentation is thorough and the architecture is sound, but you'll want to test it against your specific codebase before committing. For AI agent workflows specifically, it's worth evaluating now; for production tooling, wait for more community adoption.

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.