Atomic

A local‑first personal knowledge base that converts freeform notes into a semantically linked graph. Each piece of content—text notes, web clips, articles—is stored as an “atom” that is automatically embedded in vector space, tagged, and connected to related atoms without requiring folders or schemas. The system supports meaning‑based retrieval, allowing users to find information by concept rather than exact wording.

The platform offers AI‑augmented features such as wiki synthesis, which generates full wiki articles from tagged atoms with inline citations, and an agentic chat interface that answers questions by referencing the underlying notes. A spatial canvas visualizes the network of ideas, letting related concepts drift together, while auto‑tagging assigns metadata to each atom on entry. Integration points let external agents like Claude or Cursor access the graph.

It can run as a desktop application, a self‑hosted server, or an iOS client, synchronizing across devices through a single server that serves as the source of truth. The software is free, open source, and designed for users who want a personal, AI‑native knowledge graph without reliance on cloud services.