Atomic AI
Revolutionize RNA drug discovery with AI-driven structural biology.. [Free]
Last verified:
What is Atomic AI?
Atomic AI is a local-first, AI-native knowledge graph that transforms freeform markdown notes into a semantically connected, AI-augmented knowledge base. The fundamental unit is an
Atomic AI pricing
Pricing model: Freemium
Free and open source. No paid tiers. The desktop app is free for local personal use. Self-hosted server is free to run on your own hardware. iOS app is free. All features including semantic search, wiki synthesis, agentic chat, spatial canvas, auto-tagging, MCP integration, and API access are included at no cost. Users only pay for their own AI model/API costs if using external embedding or LLM providers.
Atomic AI pros
- Open source and completely free to use
- Local-first design keeps data on your hardware
- Semantic search finds ideas by meaning not keywords
- Wiki synthesis generates cited articles from tagged notes
- Agentic chat cites sources instead of hallucinating
- Auto-tagging eliminates need for manual folder organization
- Spatial canvas visualizes knowledge as force-directed graph
- MCP integration connects to Claude and Cursor directly
- Import Obsidian vaults in bulk
- RSS and URL ingestion automates web page capture
- Self-hosted server enables sync across all devices
- iOS app provides mobile reading and writing
- Browser extension for web clipping
- REST API with OpenAPI documentation for integrations
- No schemas or maintenance required
Atomic AI cons
- Requires self-hosting for multi-device sync
- Desktop app limited to local single-user use
- Auto-tagging depends on configured AI provider
- Embedding processing happens asynchronously may delay search
- iOS app requires self-hosted server connection
- AI provider configuration needed for full features
- Learning curve for knowledge graph concepts
- Web clipper availability depends on browser extension
Frequently asked questions about Atomic AI
What is Atomic?
Atomic is an AI-native knowledge graph that's yours end-to-end. It transforms freeform markdown notes called atoms into a semantically connected knowledge base with semantic search, wiki synthesis, and agentic chat. It is completely open source and can run as a desktop app or self-hosted server.
What is an atom?
Atoms are the fundamental unit of your Atomic knowledge base. Each atom is a markdown note containing content, title, snippet, optional source URL, capture origin, published date, hierarchical tags, and pipeline status. Atoms can be searched, tagged, linked, cited, and visualized in the graph.
How does semantic search work?
Atomic embeds every atom into vector space. When you type a question, it returns ideas based on meaning even when your notes don't use the exact words you searched for. This lets you find notes from months ago without remembering their titles.
What is wiki synthesis?
Wiki synthesis generates full wiki articles from every note, clip, and saved page under any tag. Articles include inline citations back to the source atoms and regenerate automatically as you add more content to that tag.
How does agentic chat work?
Agentic chat lets you converse with your knowledge base using RAG. It cites the specific atoms it uses for answers instead of inventing information, ensuring responses are grounded in your actual notes.
What deployment options are available?
Atomic has two common setups: a desktop app (local-first Tauri app with bundled atomic-server sidecar) for personal local use, and a self-hosted server (headless atomic-server with optional web frontend) for remote access, mobile use, browser clipping, and MCP over HTTP.
Can I import from Obsidian?
Yes, Atomic supports bulk import of Obsidian vaults. You can migrate your existing notes directly into Atomic as atoms.
How does auto-tagging work?
When you create or update an atom, an asynchronous pipeline analyzes the full content using an LLM to extract structured hierarchical tags. This happens in the background while you keep writing, and no folders are needed.
What is the MCP integration?
MCP Integration lets you plug Atomic into Claude or Cursor. Your notes become available in your AI agent's hands through the MCP protocol, enabling agents to use your knowledge base directly.
Is Atomic free?
Yes, Atomic is completely free and open source. There are no paid plans or tiers. You own your notes, hardware, and model choices end-to-end with no subscription fees.