LibreAI / Produtividade / obsidian-mind
// productivity · breferrari/obsidian-mind

obsidian-mind

obsidian-mind é um projeto de produtividade de código aberto de breferrari (breferrari/obsidian-mind no GitHub). Tem 4.955 estrelas no GitHub e 556 forks, é publicado sob a licença MIT e foi atualizado pela última vez no GitHub em 6 de outubro de 2026.

A self-organizing Obsidian vault that gives AI coding agents persistent memory across sessions.

★ 5,0 mil estrelas⑂ 556 forksMITatualizado out 2026
▦ visual: o README deste projeto não inclui nenhuma captura incorporável,
então aqui vai um marcador. O guia abaixo está completo mesmo assim.
// dados do projeto

Dados do projeto

Estrelas no GitHub
4.955
Forks
556
Licença
MIT
Última atualização no GitHub
6 de outubro de 2026
Categoria
Produtividade
Tipo de ficha
Guia de instalação completo
// como funciona

Como funciona

  1. Clarify what your agent forgets

    Note what you re-explain every session — project status, decisions, people, preferences — that a persistent vault should hold.

  2. Create your vault

    Create a new Obsidian vault and install obsidian-mind into it with shardmind (or clone the repo).

  3. Set up semantic search

    Install QMD and run the vault's bootstrap so the agent can search notes by meaning.

  4. Start your first session

    Tell the agent 'start session', brain-dump your day with the provided commands, and wrap up so everything is filed and linked.

// recursos

Por que as pessoas gostam tanto

Persistent agent memory

Sessions begin from your North Star, active projects, open tasks and recent memories instead of a blank slate.

Self-organizing capture

Brain dumps are filed automatically — people notes updated, decision records created, wins added to your brag doc, project notes amended.

Semantic recall with QMD

Local semantic search finds what you decided even when the note is titled differently, and brain/ guidance topics load on demand.

Tiered, token-efficient loading

Always-loaded context is budget-capped, specific context loads on demand via search, and subagents get sharper context — the vault is never dumped wholesale into the prompt.

Workflow commands included

Ships session rituals such as /om-standup, /om-dump, /om-incident-capture and a wrap-up that verifies links, updates indexes and spots uncaptured wins.

Works across agent CLIs

Designed for Claude Code, Codex CLI and Gemini CLI, with an optional MCP wiring so the vault is available in every directory.

Advertisement
// execute você mesmo

Execute você mesmo

  1. Clone the repository

    Clone obsidian-mind, or install it into a fresh vault with shardmind as shown in the next step.

    git clone https://github.com/breferrari/obsidian-mind.git
    cd obsidian-mind
  2. Install into your vault

    Create a vault directory and install the obsidian-mind system into it.

    npm install -g shardmind
    mkdir my-vault && cd my-vault
    shardmind install github:breferrari/obsidian-mind
  3. Enable semantic search

    Install QMD and bootstrap the vault index, then query it.

    npm install -g @tobilu/qmd
    node --experimental-strip-types .scripts/qmd-bootstrap.ts
    qmd --index obsidian-mind query "what did we decide about caching"
// casos de uso

O que você pode criar

Advertisement
// perguntas frequentes

Perguntas frequentes

Who maintains obsidian-mind?

obsidian-mind is maintained by breferrari and contributors.

Is it free?

Yes — it is MIT licensed. You need Obsidian (free for personal use) and an AI coding agent you already use.

Does it need a paid API key?

The vault and QMD search run locally; the agent CLI you pair it with (Claude Code, Codex, Gemini) uses your existing account with that tool.

How does it compare to an agent's built-in memory?

Built-in memories are opaque and tool-specific; obsidian-mind keeps memory as your own linked Markdown notes in Obsidian — readable, editable, portable across agent CLIs, and semantically searchable.

How do I get started?

Create a vault, run shardmind install github:breferrari/obsidian-mind, bootstrap QMD, and tell your agent to start a session.

// continue explorando

Ferramentas semelhantes

Produtividade

obsidian-second-brain

Persistent memory for Claude Code and other CLI agents, stored as plain Markdown in your Obsidian vault

★ 4,7 mil estrelasMIT
Ver o guia →
Produtividade

jarvis-dashboard

An Obsidian dashboard for your Claude Code sessions.

★ 102 estrelasMIT
Ver o guia →
Produtividade

CowAgent

Open-source personal AI assistant & Agent Harness. Plans tasks, runs tools and skills, self-evolves with memory and knowledge. Multi-agent, multi-model, mult…

★ 47,2 mil estrelasMIT
Ver o guia →
Produtividade

screenpipe

Record everything you do on your computer - searchable, private, and ready for agents

★ 21,9 mil estrelasNOASSERTION
Ver o guia →
Produtividade

rowboat

An AI coworker with long-lived memory and multiplayer collaboration spaces

★ 18,0 mil estrelasApache-2.0
Ver o guia →
Produtividade

LokalBot

Local LLM workhorse that keeps a memory of your workday.

★ 48 estrelasGPL-3.0
Ver o guia →
// alternativas

Alternativas em Produtividade

Os outros projetos de Produtividade com mais estrelas no catálogo do LibreAI — a lista realista de alternativas se obsidian-mind não for o ideal para você.

// mais em productivity

Mais em Produtividade

Projetos de Produtividade próximos no catálogo, os mais próximos em estrelas de obsidian-mind: continue explorando dos dois lados dele.

Baixe o obsidian-mind

Grátis e aberto: pegue o código-fonte, dê uma estrela no repositório e comece a criar.