second brain

Personal knowledge base for research, projects, ideas, and business. Inspired by karpathy's LLM wiki system.

Structure

Folder	Purpose
sources/	Raw ingested documents — papers, articles, notes, data
entities/	Pages about people, organizations, tools, products
concepts/	Definitions, mental models, technical concepts
projects/	Active and past projects
business/	Business ideas, strategy, market research
ideas/	Raw ideas not yet classified

Key Files

• index.md — Content catalog, searchable overview of everything in the wiki
• log.md — Append-only chronological ingest/query/lint log
• schema.md — Conventions, workflows, and maintenance rules

Workflows

Ingest a new source (small, ≤ ~20 pages)

1. Drop the file into sources/
2. Open a conversation with an LLM, paste schema.md + the source
3. Ask it to: summarize → create/update entity or concept pages → update index.md → append to log.md

Ingest a large PDF (any size)

ANTHROPIC_API_KEY=sk-... python3 ingest.py sources/my-paper.pdf

ingest.py extracts all text from the PDF, splits it into 30-page chunks (configurable with --chunk-pages), calls Claude on each chunk to preserve all details, then merges everything into wiki pages automatically. Outputs: sources/<slug>.md, relevant concepts/ and entities/ pages, an index.md entry, and a log.md line.

Query

1. Open a conversation, load relevant pages from concepts/ or entities/
2. Ask your question — instruct the LLM to cite page names
3. If the answer is worth keeping, save it as a new page

Lint (periodic health check)

• Ask the LLM to scan for contradictions, orphan pages, stale claims, and missing cross-references
• Log the lint pass in log.md