Agents365-ai

rak — Semantic and hybrid search over your Zotero library, powered by local embeddings

11
1
100% credibility
Found Mar 24, 2026 at 11 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
Python
AI Summary

rak is a desktop tool that adds smart natural-language search and AI question-answering to personal Zotero research libraries.

How It Works

1
📚 Discover rak

You hear about rak, a helpful companion for your Zotero collection of research papers, making it easy to find what you need.

2
💻 Add rak to your setup

You simply add rak to your computer in a quick step, ready to work with your existing paper organizer.

3
🔄 Teach rak your papers

You tell rak to scan and remember all your saved papers, creating a smart map of everything in moments.

4
🔍 Search like magic

You type a question in everyday words, and rak instantly shows the most relevant papers with key highlights.

5
Pick your way to explore
Quick search

See matching papers and snippets right away to grab what you need fast.

💬
Chat with helper

Start a back-and-forth talk where rak answers questions using your papers.

🎉 Master your library

Now you effortlessly find, understand, and connect ideas across your entire research collection.

Sign up to see the full architecture

4 more

Sign Up Free

Star Growth

See how this repo grew from 11 to 11 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 zotero-rag-cli?

Rak is a Python CLI for semantic and hybrid search over your Zotero library, powered by local embeddings. It indexes papers from Zotero storage—including PDF full text into chunks—letting you query with natural language like "cell fate determination" to get relevant titles, scores, and matching snippets via `rak search` or `rak --json search`. Pair it with `rak ask` or `rak chat` for LLM-powered Q&A, all running locally without sending data off-machine.

Why is it gaining traction?

It stands out with hybrid search blending semantic vectors and BM25 keywords via RRF fusion, returning precise snippets instead of whole papers, plus incremental indexing and export to CSV/BibTeX. The MCP server integrates seamlessly with LM Studio or Cursor for tool calls, and it supports swapping embedding models or providers like Ollama/OpenAI. Fully local operation ensures privacy, beating cloud-only alternatives.

Who should use this?

Academics or researchers managing large Zotero libraries for lit reviews—run `rak index` once, then query during writing or feed JSON results to Claude Code for summaries. It's ideal for those pairing Zotero with local LLMs, avoiding API costs while filtering by collections/tags.

Verdict

Grab it if you need fast RAG over Zotero (install via `uv tool install zotero-rag-cli`), but note the 1.0% credibility score and 11 stars signal early beta status—solid docs and CLI polish make it worth trying, just test on a subset first.

(187 words)

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.