Browse / LLM wiki / Second brain / llm-wiki

llm-wiki

shashankGS10

A local, agent-driven knowledge base using plain Markdown files and Claude Code — no vector databases, no embeddings, no complex RAG pipelines.

Open shashankGS10/llm-wiki → Checked against the four rules.

Built for one agent

ships CLAUDE.md and nothing for other agents.

In its own words

README.mdshashankGS10/llm-wiki

A local, agent-driven knowledge base using plain Markdown files and Claude Code — no vector databases, no embeddings, no complex RAG pipelines.

Based on Andrej Karpathy's LLM Wiki concept.

How It Works

  • Drop raw content (transcripts, articles, notes) into inbox/
  • Run claude in this directory and ask it to ingest
  • Query your knowledge base by asking questions — the agent reads files, follows wikilinks, and synthesizes answers

Demo

See the Loom walkthrough: [link TBD]

The whole README. Open the repo ↗

Read from the repository's git tree on 2026-08-17.