IgorWarzocha

IgorWarzocha / pi-semantic-grep

Public

Pi extension that enables agents to look things up via natural language query.

18
0
100% credibility
Found Apr 28, 2026 at 18 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
TypeScript
AI Summary

An extension for the pi coding agent that provides semantic search to locate relevant code and documentation by conceptual meaning using AI embeddings.

How It Works

1
📖 Discover smarter code search

While using your coding companion pi, you hear about a helpful add-on that finds code and docs by meaning, not just exact words.

2
🛠️ Add the helper to pi

You simply add this tool to your pi setup with an easy install command.

3
🧠 Connect your AI service

You point the tool to your preferred AI service that understands the meaning behind text.

4
🌟 Start pi in your project

When you open pi in your project folder, it automatically creates a smart map of your code and documents behind the scenes.

5
❓ Ask about code concepts

Your pi agent now has a special search tool to query things like 'Where is user login handled?' or 'How are prompts built?'

6
🔍 See ranked matches

It instantly shows the top relevant code sections with file locations, previews, and confidence scores.

🎉 Unlock project insights

You now effortlessly discover and understand code across files, making development faster and more intuitive.

Sign up to see the full architecture

5 more

Sign Up Free

Star Growth

See how this repo grew from 18 to 18 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 pi-semantic-grep?

Pi-semantic-grep is a TypeScript extension for the pi coding agent that enables agents to look things up via natural language query using semantic grep. Instead of literal text matching, it indexes your repo's code and docs into a local SQLite database with OpenAI-compatible embeddings, delivering ranked matches by meaning—like "where is auth handled?"—with file paths, snippets, and scores. Install via `pi install npm:@howaboua/pi-semantic-grep`, tweak config for your local embeddings server, and agents call `semantic_grep({query: "prompt building code"})`.

Why is it gaining traction?

It stands out by giving pi agents semantic search over exact grep, perfect when keywords fail, unlike github copilot chat extensions that stick to syntax. Incremental indexing at session start keeps it fast and repo-local, no cloud or heavy vector DBs needed, and safety rules block indexing home dirs or non-projects. Developers hook on the compact results that expand in pi, bridging natural language queries to precise code locations.

Who should use this?

Pi users building agents for repo exploration, like AI coding assistants querying "tool dispatch logic" across files. Backend devs tracing session flows, or docs-heavy teams finding examples without manual hunts. Skip if you're in vscode with github copilot extensions already covering basics.

Verdict

Early maturity with 18 stars and 1.0% credibility score means test in a side project first—docs are solid, config flexible, but expect tweaks for large repos. Worth trying if pi agents need semantic lookup; pair with local embeddings like LM Studio for quick wins.

(198 words)

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.