entireio

entireio / pgr

Public

PGR is an experimental, stateless MCP code-search server for studying how ranking, latency, and output shaping affect agentic search.

mcp-server research
16
1
100% credibility
Found May 07, 2026 at 17 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
Python
AI Summary

pgr is an experimental server that enhances code search for AI agents by ranking results to prioritize implementation files over tests and noise.

How It Works

1
🔍 Discover Smart Code Search

You hear about a helpful tool that makes AI assistants better at finding code in projects.

2
📥 Get the Tool

You download and set up the tool on your computer with a simple install.

3
📁 Pick Your Project

You choose a folder with code you want your AI to explore.

4
🔗 Link to Your AI

You connect the tool to your AI assistant so it can use smart search right away.

5
💬 Ask About the Code

You tell your AI to search for something specific, like where a feature is built.

6
⭐ See Better Results

Your AI gets the most useful files first, skipping tests and extras.

✅ AI Works Faster

Your AI helper navigates code projects quicker and more accurately every time.

Sign up to see the full architecture

5 more

Sign Up Free

Star Growth

See how this repo grew from 17 to 16 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 pgr?

pgr is an experimental Rust MCP server for agentic code-search, wrapping ripgrep to rank and shape results for AI workflows. It prioritizes implementation files over tests, formats snippets for quick model decisions, and offers stateless tools like search_code, read_code, find_files, and list_dir over stdio—no indexes or daemons needed. Developers get faster paths to relevant code, studying how ranking, latency, and output affect agent behavior.

Why is it gaining traction?

It stands out by de-noising ripgrep output: definitions first, source over tests, with agent-tuned formatting that cuts search loops. Public benchmarks compare it to baselines and indexed alternatives like fff, showing better first-query hits and fewer reads. MCP stdio makes it plug-and-play for any client, no setup overhead.

Who should use this?

AI engineers building agentic tools for repo analysis, like checkpoint debugging in projects such as entireio/cli. Researchers tweaking code-search for LLMs, or backend devs needing ranked retrieval in MCP pipelines without state.

Verdict

Try pgr for agentic experiments—cargo install, set cwd to your repo, and query away. At 16 stars and 1.0% credibility, it's a raw research artifact with strong docs and repro benchmarks, but expect tweaks for production.

(198 words)

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.