GitHub Repository

Funes is a Git-based framework for LLM-managed knowledge work: an AI Librarian ingests raw sources, builds an interlinked Markdown knowledge base, and uses it to produce cited reports, analyses, and other outputs.

8 stars

A self-organizing digital Zettelkasten for research

Name: A self-organizing digital Zettelkasten for research
Availability: InStock
Author: bethanyhunt

by bethanyhunt·May 27, 2026·1 point·0 comments

Visit Project View on HN

AI Analysis

●●SolidBig BrainCozy

Git-based AI knowledge base with health checks beats cloud-only alternatives.

Strengths

•Plain-text Git workflow ensures portability and versioning forever
•Health check system maintains knowledge base coherence over time
•Example library shows concrete outputs before you commit to using it

Weaknesses

•Competes with Obsidian and Logseq with AI plugins already available
•Requires agentic coding tool setup, not a standalone application

Post Description

I have tried starting both analog and digital Zettelkastens several times but never kept up with them because the maintenance always seemed overwhelming. The "LLM Knowledge Bases" Karpathy tweet made me wonder if an AI-assisted version might work, and it worked beyond my expectations, so I thought others might find it useful.

You dump your sources in the raw folder and ask your AI of choice to ingest/process them. This produces a linked wiki of cited summaries, topics, and concepts. Then you can query for other kinds of outputs based on your sources, like research summaries, answers to questions, gap analyses, etc. You can later add additional sources to raw to be processed and integrated into your library.

I made a mini sample library so that you can see the kinds of outputs it produces. Browse the wiki and outputs folders in starter-library: https://github.com/ulyssestenn/funes-example-ai/tree/main/st...

Would be interested in your feedback.