catlog22

catlog22 / codexlens-search

Public

Lightweight semantic code search engine — 2-stage vector + FTS + RRF fusion + MCP server for Claude Code

20
2
100% credibility
Found Mar 23, 2026 at 20 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
Python
AI Summary

A semantic code search engine that integrates with AI assistants like Claude via a simple configuration, enabling hybrid searches across vector embeddings, full-text, code structure, and regex patterns.

How It Works

1
🔍 Discover a smart code helper

You hear about a tool that lets your AI assistant instantly find and understand code in your projects.

2
📦 Get it set up easily

Download and add a simple note to your project so your AI knows about this helper.

3
🔗 Connect your AI

Link it to your favorite AI service with a quick setting, like giving it a thinking brain.

4
🚀 Start it up

Turn it on once, and it watches your files automatically.

5
📂 Build your code map

Tell your AI to scan your project, and it creates a smart map of all your code in seconds.

6
💭 Ask questions naturally

Chat with your AI using everyday words like 'show me the login function' and get exact matches.

✅ Code insights at your fingertips

You now navigate huge projects effortlessly, saving hours and boosting your productivity.

Sign up to see the full architecture

5 more

Sign Up Free

Star Growth

See how this repo grew from 20 to 20 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 codexlens-search?

Codexlens-search is a Python-based semantic code search engine that plugs into Claude Code via an MCP server, letting you query codebases with natural language, symbols, or regex. Drop it in with pip install and a .mcp.json config—supports local models or OpenAI/Jina APIs, with hybrid results from vectors, full-text search, AST graphs, and ripgrep. Handles 10s cold-start searches by grepping relevant files first, then indexing on-demand.

Why is it gaining traction?

It skips bloated RAG stacks for a lightweight semantic search setup that runs offline with tiny models (33MB default) and accelerates on any Windows GPU via DirectML—no CUDA hassle. File watching auto-reindexes changes, and CLI tools like sync/status make it dead simple for quick codebase navigation. Devs dig the no-wait hybrid fusion that blends semantic relevance with exact matches.

Who should use this?

Claude Code users indexing Python, JS, or Go repos for fast symbol/ref lookups. Solo devs or small teams searching large monorepos without cloud costs. Anyone tired of grep/dash scrolling through noisy results in VS Code or Cursor.

Verdict

Grab it for lightweight semantic search if you're in Claude's ecosystem—installs clean, docs solid with benchmarks, MIT licensed. At 20 stars and 1.0% credibility, it's early but battle-tested on its own codebase; production-ready for prototypes, watch for wider adoption.

(178 words)

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.