yologdev

The Self-Growing Karpathy LLM Wiki — grown by an AI agent yoyo from Karpathy's founding prompt

17
2
100% credibility
Found Apr 07, 2026 at 17 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
TypeScript
AI Summary

A web app for creating and managing a personal wiki by ingesting sources, querying content, and maintaining links using language models, demonstrated through an autonomous AI agent that builds and improves it over time.

How It Works

1
🔍 Discover the wiki builder

You stumble upon this fun project online that promises a personal knowledge collection powered by smart helpers.

2
🚀 Get it running at home

Follow simple steps to start the web app on your own computer so you can begin building right away.

3
🤖 Link a thinking helper

Connect a smart AI service that makes your wiki understand and connect your notes automatically.

4
📄 Add your first notes

Paste in text from articles or links from the web, and watch it turn into a neat wiki page.

5
Explore your growing wiki
🔗
Browse pages

Wander through linked pages to see how ideas connect.

Ask questions

Type a question and get answers pulled straight from your notes.

🧹
Check health

Run a quick scan to fix loose ends or gaps in your wiki.

🌟 Your smart wiki thrives

Sit back as your personal knowledge base grows smarter, more connected, and always ready to help.

Sign up to see the full architecture

4 more

Sign Up Free

Star Growth

See how this repo grew from 17 to 17 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 karpathy-llm-wiki?

This TypeScript web app, built with Next.js, turns raw sources into a personal LLM-powered wiki. Paste a URL or text, and it ingests content—fetching pages, summarizing via LLMs, and creating interlinked markdown pages with summaries, key points, and concepts. Query for cited answers, browse a graph of connections, or lint for orphans, contradictions, and missing links, all stored persistently on disk as an anti-RAG knowledge base.

Why is it gaining traction?

It's the self-growing Karpathy LLM wiki, bootstrapped from Karpathy's founding prompt by an AI agent called yoyo that commits production code every four hours—no human coding after the baseline markdown file. Developers dig the live demo of agent-driven evolution: watch it add features like query saving or graph views autonomously, with journals tracking progress. The harness enforces builds, tests, and reverts, proving reliable self-directed dev from one prompt.

Who should use this?

AI researchers or devs building personal wikis from blog posts, papers, or notes, especially Karpathy fans wanting an interlinked hub without manual linking. Indie hackers prototyping LLM tools who need quick ingest/query/lint cycles. TypeScript/Next.js users experimenting with agent workflows for side projects.

Verdict

Grab it if agent hype excites you—spin up locally with an Anthropic or OpenAI key and ingest your first source in minutes. At 17 stars and 1.0% credibility, it's raw and experimental (light tests, early docs), but the yoyo's autonomous growth makes it a must-watch for LLM agent pioneers.

(198 words)

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.