s2b-dev/smart-second-brain

Smart Second Brain – An AI‑enhanced Obsidian Plugin

What it is – A free, open‑source plugin for the Obsidian note‑taking app that adds AI‑powered search, a dynamic knowledge‑graph view, and an interactive chat assistant that can read and write your vault’s markdown files.

Why it matters – Traditional Obsidian search is purely keyword‑based and the built‑in graph only shows explicit links. Smart Second Brain augments both with semantic embeddings and a lightweight “agent” that can answer questions using the content of your notes, generate charts, and even modify files on your behalf.

Key features

• Hybrid search – Combines keyword matching with vector‑based meaning similarity. Works out‑of‑the‑box with plain keyword search; adding an embedding model unlocks the semantic layer.
• Smart Graph – Automatically clusters notes into topics derived from links, tags, and the actual text, giving you a higher‑level view of what your vault is about.
• Agents – A chat‑style assistant that can read your notes, create dashboards/widgets, and write back to the vault. It supports memory, skills, and can be extended via “skills” or custom agents.
• Privacy‑first – No data leaves your device unless you explicitly configure an external embedding model or LLM provider. The only network call is an optional daily check for plugin updates.

How it works

1. Search & Graph – By default they use simple keyword techniques. When you supply an embedding model (e.g., a local sentence‑transformer), the plugin builds vector representations of each note and performs similarity search; the same embeddings feed the topic‑clustering for the Smart Graph.
2. Agent – Once you configure an LLM provider (OpenAI, Anthropic, etc.), the plugin sends the relevant note excerpts to the model, receives a response, and can act on it (e.g., insert a new note, update a table, generate a chart). The architecture is described in AGENTS.md.

Getting started

• Install from Obsidian’s Community Plugins marketplace and enable it – the search and graph work immediately.
• To unlock semantic features, add an embedding model in the plugin settings (local or remote) and select it for both search and graph.
• Add an AI provider to enable the chat assistant.

Community & contribution

• Issue tracker for bugs/feature requests, discussion forums for Q&A and sharing custom agents, and a CONTRIBUTING guide that even covers AI‑assisted development.
• Licensed under MIT, so you can fork, modify, or embed it in your own workflows.

Bottom line – Smart Second Brain brings state‑of‑the‑art AI capabilities directly into the Obsidian ecosystem while keeping data local, making it a practical tool for anyone who wants a more intelligent personal knowledge base.