GrepAI iconGrepAI

oss Free Star1k

Privacy-first semantic code search CLI that runs 100% locally with Ollama, providing meaning-aware queries and call graphs for AI coding agents

1.7K GitHub Stars
27% Token Savings
100% Local Processing

Overview

GrepAI is a privacy-first semantic code search CLI designed to replace traditional keyword-based search with meaning-aware queries. Unlike ripgrep or grep, GrepAI indexes the semantic meaning of your code, letting developers and AI agents find relevant code by what it does rather than just text matches. The tool runs 100% locally using Ollama for embeddings—no cloud, no API keys, no data leaving your machine. With MCP server integration, GrepAI works directly with Claude Code, Cursor, and Windsurf, reducing token costs by up to 27% by providing more precise context retrieval.

The Verdict

Who Should Use GrepAI?

Best For

  • Developers using AI coding assistants who want to reduce token costs
  • Privacy-conscious teams who can't send code to cloud services
  • Users of Claude Code, Cursor, or Windsurf seeking better context
  • Developers navigating large, unfamiliar codebases
  • Anyone frustrated with keyword-only code search

Not Ideal For

  • Teams needing enterprise features like cross-repo search
  • Developers without local GPU/CPU for Ollama embeddings
  • Simple projects where ripgrep is sufficient
  • Those requiring cloud-hosted or managed solutions

What's Great

  • 100% local processing—no data leaves your machine
  • MCP server works natively with Claude Code, Cursor, Windsurf
  • Semantic search finds code by meaning, not just keywords
  • Call graph tracing for function relationship analysis
  • File watcher auto-updates index as you code
  • Single binary, no dependencies, indexes 10K files in seconds

Watch Out For

  • Requires Ollama, LM Studio, or OpenAI for embeddings
  • Initial indexing requires CPU/GPU resources
  • Limited to supported languages (Go, TS, Python, etc.)
  • No cross-repository or team-wide search

Pricing

View all features & details

Key Features

  • Semantic code search by meaning
  • Call graph tracing
  • MCP server integration
  • Auto-updating file watcher
  • Millisecond search latency

Supported Languages

  • Go
  • TypeScript / JavaScript
  • Python
  • PHP, Java, C#, C++
  • Rust, Zig

Embedding Providers

  • Ollama (default, local)
  • LM Studio (local)
  • OpenAI (cloud, optional)

Integrations

  • Claude Code
  • Cursor
  • Windsurf
  • Any MCP-compatible tool

How It Compares

Feature GrepAI Codemogger Sourcegraph
Search Type Semantic + Call Graphs Semantic + Keyword AI + Deterministic
Privacy 100% Local 100% Local Cloud / Self-hosted
MCP Server Yes Yes Yes
Language Go (C bindings) TypeScript Multi-language
Best For Privacy-first semantic search Simple local indexing Enterprise teams

User Reviews

Loading reviews...