Atomic

Atomic is an open-source knowledge management tool that collects notes, articles, and research into an AI-augmented knowledge graph. It automatically embeds, tags, and links content to related ideas, allowing your knowledge base to organize itself as it grows. Available as a macOS desktop app, self-hosted server, iOS app, browser extension, and MCP server, Atomic runs wherever you work and keeps your data under your control.

• Semantic Search — Find ideas by meaning using vector embeddings that surface related notes even when exact keywords don't match.
• Wiki Synthesis — LLM-generated wiki articles synthesize everything under a tag, with inline citations linking back to original source notes; articles update incrementally as you add more content.
• Agentic Chat — Chat with an AI that searches your notes mid-conversation, scoped to a single tag or your whole library, citing sources rather than hallucinating.
• Spatial Canvas — Visualize your knowledge as a force-directed graph where semantically related atoms cluster together; zoom, pan, and explore the topology of your ideas.
• Auto-Tagging — Drop in a note and receive a tag tree covering topics, people, places, organizations, and events — no manual folder maintenance required.
• MCP Integration — Give Claude, Cursor, or any MCP client direct access to your knowledge base to search, read, and create atoms without leaving your workflow.
• Atoms — Every piece of knowledge is an atom — a markdown note, saved article, or web clip — automatically tagged, embedded, and connected; pull content in from URLs, RSS feeds, and a browser extension.
• Self-Hostable — Run Atomic as a headless server on your own infrastructure so your data stays entirely yours.