Obsidian Second Brain is a persistent memory system for AI agents that stores context as plain markdown in your Obsidian vault.
The tool solves the problem of AI agents losing context between sessions by maintaining a shared knowledge base accessible to multiple CLI agents. It works by storing project decisions, people, and architectural notes in your Obsidian vault, which agents can query and update. The system includes hybrid semantic search to retrieve relevant information, self-rewriting notes that update automatically, and scheduled agents that maintain the vault without manual intervention.
You should adopt this tool if you work with multiple AI agents across different platforms and want to avoid re-explaining your project context in every session. It suits teams and individuals who use Claude Code, Grok Bot, Codex, Gemini, or other supported CLI agents and want a unified memory layer that persists across tool switches. The vault itself becomes the source of truth, outliving whichever agent platform you use at any given time.
The project shows active development with a substantial command set and cross-platform support spanning multiple agent ecosystems. The tool includes web research capabilities without requiring API keys and offers scheduled automation features that run independently. The codebase is written in Python and maintains a public homepage with documentation separate from the repository.
