Back to Projects

CodeLens

AI-powered GitHub code exploration tool with semantic search, automated documentation, and natural language codebase querying.

Summary

A developer productivity tool that understands large repositories using LLMs, pgvector, and type-safe APIs.

Technologies

Next.js
TypeScript
PostgreSQL
Ppgvector
TtRPC
CClerk
Redis
GGitHub API

Problem

Developers spend hours jumping across files, reading unfamiliar code, and manually searching large repositories. Traditional search is keyword-based and doesn’t understand code semantics, relationships, or intent.

Solution

Built an AI-powered analysis engine that lets developers query codebases in natural language, generate documentation, and perform semantic code search using pgvector embeddings and LLM reasoning.

Why This Project

I built CodeLens because I constantly deal with large, messy codebases in real-world projects, and context-switching kills productivity. This tool reduces research time from hours to minutes by making the codebase itself ‘talk back.’