benmaster82

benmaster82 / Kwipu

Public

Ask questions across your Markdown notes using a fully local Graph RAG engine. Built for Obsidian vaults, works with any folder of Markdown files. Extracts entity-relation triples from wikilinks & YAML frontmatter, retrieves answers via hybrid search (vector + BM25 + temporal). Multilingual. No cloud. Runs on Ollama.

embeddings graph-rag help-wanted knowledge-graph llama-index
43
6
100% credibility
Found May 03, 2026 at 43 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
Python
AI Summary

Kwipu is a local application that builds a knowledge graph from markdown notes in a folder, enabling natural language queries with hybrid retrieval to connect and retrieve information across files.

How It Works

1
🔍 Discover Kwipu

You hear about Kwipu, a helpful tool that connects all your notes so you can ask questions and get smart answers from them.

2
📁 Gather your notes

Put your everyday notes, like from your journal or Obsidian app, into one simple folder on your computer.

3
🚀 Start your knowledge helper

Launch the program, and it quietly reads all your notes to build a map of how everything connects.

4
⏳ Let it learn

The first time, it takes a few minutes to understand your notes fully, then it's ready forever.

5
💬 Ask any question

Type a question in plain words, like 'What did I learn about projects last month?' and watch it think.

6
👀 See connected answers

Get clear responses that pull together info from different notes, with sources listed so you trust it.

🎉 Your notes come alive

Now your knowledge base answers questions instantly, updates when you add notes, and helps you remember everything effortlessly.

Sign up to see the full architecture

5 more

Sign Up Free

Star Growth

See how this repo grew from 43 to 43 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 Kwipu?

Kwipu turns any folder of Markdown files—especially Obsidian vaults—into a queryable knowledge graph you can ask natural language questions against, like "ask github repo questions" but fully local. Drop your notes in a folder, fire up Python and Ollama, and it extracts entities from wikilinks and YAML frontmatter to connect info across files via hybrid search (vector, keyword, temporal). Answers stay grounded in your docs with source citations, no cloud needed, and it auto-syncs on file changes.

Why is it gaining traction?

It stands out for Obsidian users tired of manual searching scattered notes, delivering multilingual hybrid retrieval that's faster in fast mode and hallucination-proof via strict prompts. Real-time folder watching means edits reflect instantly without rebuilds, and you can build once with a beefy model then query with a lightweight one. Devs digging local RAG dig the "ask questions to ai" vibe without GitHub Copilot's telemetry or "quipu mayo"-style knot-tracking for knowledge.

Who should use this?

Obsidian power-users managing project notes, research vaults, or multilingual docs who want to ask questions like "what's the status of the Q3 deliverables?" across files. Python devs building personal knowledge bases for code snippets, meeting recaps, or bible-study notes. Teams avoiding cloud RAG for sensitive info, especially with Ollama setups.

Verdict

Worth a spin for local-first note querying if you have 16GB RAM and patience for initial builds (minutes to hours on CPU), but at 23 stars and 1.0% credibility, it's early-stage—solid docs, no tests visible, so tweak configs yourself. Fork-friendly MIT for production tweaks.

(198 words)

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.