explaingit

andresbuonaiuto/cerebro

Analysis updated 2026-05-18

1PythonAudience · vibe coderComplexity · 2/5LicenseSetup · moderate

TLDR

Cerebro turns your books and PDFs into a linked markdown knowledge base that an AI keeps updated, so you get cited, traceable answers instead of guesses.

Mindmap

mindmap
  root((cerebro))
    What it does
      Builds linked markdown wiki
      Keeps knowledge current
      Cites sourced answers
      Flags contradictions
    Tech stack
      Python
      Markdown
      Obsidian
      Claude skill
    Use cases
      Personal knowledge base
      Idea validation
      PDF to markdown conversion
    Audience
      Vibe coders
      Researchers
      Idea validators

Code map

Detail Auto

An interactive map of this repo's files and how they connect — its source is parsed live in your browser. Click Visualize to build it.

filefunction / class

What do people build with it?

USE CASE 1

Build a growing, cited knowledge base from your books and PDFs that an AI keeps current.

USE CASE 2

Validate business or research ideas against your sources and get a cited supports or contradicts answer.

USE CASE 3

Convert PDFs to clean markdown first to cut the cost of having an AI read them.

What is it built with?

PythonMarkdownObsidianClaude

How does it compare?

andresbuonaiuto/cerebro0xustaz/streamgatea-bissell/unleash-lite
Stars111
LanguagePythonPythonPython
Setup difficultymoderatehardhard
Complexity2/54/54/5
Audiencevibe coderdeveloperresearcher

Figures from each repo's GitHub metadata at analysis time.

How do you get it running?

Difficulty · moderate Time to first run · 30min

Best used with Obsidian for browsing, and a Claude skill install for the smoothest workflow.

Use freely for any purpose, including commercial use, as long as you keep the copyright notice.

In plain English

Cerebro turns books, PDFs, and other documents into a persistent knowledge base that an AI keeps up to date, so you can ask questions and get answers backed by real citations instead of the AI reinventing an answer from scratch every time. Many document chat tools rediscover the same files for every question, so nothing builds up. Cerebro instead has the AI read each source once, turn it into a set of linked markdown pages, and keep updating those pages as new material is added. The idea is that you gather the sources and ask the questions, while the AI does the reading, summarizing, cross referencing, and organizing. You then browse the resulting notes using Obsidian, a popular note taking app for viewing linked markdown files. There are four core actions: creating a new knowledge base for a topic, ingesting a source so the AI reads it and updates related pages, querying it to ask a question or check an idea and get a cited answer that admits when there is no coverage rather than guessing, and linting it to find contradictions, outdated claims, or gaps. There are a few ways to get started. With Claude, you can build and install cerebro as a skill, then simply say you want to create a knowledge base and ask it to read a document you provide. With other filesystem based AI agents like Codex, Cursor, or Gemini CLI, you copy a starter template and point the agent at the project's blueprint document. With ChatGPT, you can paste a prepared prompt into a custom assistant and upload your existing pages for question answering only. The project includes a full example knowledge base built from two classic public domain books, showing how source notes, concept pages, and cross source summaries work together, including a documented case where the two books disagree. A separate helper script converts PDFs into clean markdown first, which reduces the cost of having an AI read them, especially for scanned or image heavy documents. Cerebro is based on an idea for an LLM knowledge wiki described by Andrej Karpathy, turning that idea into a ready to use kit with confidence levels, contradiction handling, and structured answers. This project is aimed at vibe coders, researchers, and anyone who wants an AI assistant with a memory that improves over time. It is released under the MIT license.

Copy-paste prompts

Prompt 1
Help me set up cerebro as a Claude skill and create my first knowledge base from a PDF.
Prompt 2
Explain the four core operations in cerebro: create, ingest, query, and lint.
Prompt 3
How does cerebro avoid hallucinating answers when a question isn't covered by my sources?
Prompt 4
Walk me through the example business brain and how it handles a contradiction between two sources.

Frequently asked questions

What is cerebro?

Cerebro turns your books and PDFs into a linked markdown knowledge base that an AI keeps updated, so you get cited, traceable answers instead of guesses.

What language is cerebro written in?

Mainly Python. The stack also includes Python, Markdown, Obsidian.

What license does cerebro use?

Use freely for any purpose, including commercial use, as long as you keep the copyright notice.

How hard is cerebro to set up?

Setup difficulty is rated moderate, with roughly 30min to a first successful run.

Who is cerebro for?

Mainly vibe coder.

Open on GitHub → Explain another repo

This repo across BitVibe Labs

Verify against the repo before relying on details.