Atomic Place is a local-first workspace for documents, tables, files, chat and apps. It works on your device and can sync through a self-hosted AtomicServer or optional managed services. The underlying Atomic Data specification is open.

This repo includes:

• Atomic Place, the workspace in the browser, on desktop and on mobile.
• AtomicServer, the self-hostable server for storage, search, sharing and sync.
• docs documentation / specification for Atomic Data (docs.atomicdata.dev).
• Atomic Data Browser, the React front-end for Atomic Place.
• @tomic/lib JS NPM library.
• @tomic/react React NPM library.
• @tomic/svelte Svelte NPM library.
• atomic_lib Rust library.
• atomic-cli terminal client.
• flutter a Dart / Flutter client (with demo app AtomicCanvas)

Status: alpha. Breaking changes are expected until 1.0.

• 🏠 Local-first: works offline in the browser, syncs when you reconnect with Atomic Sync.
• 🔄 Real-time collaboration: live cursors, typing indicators, and following what a teammate is doing.
• 📄 Documents: collaborative rich text, like Google Docs or Notion.
• 🗄️ Tables: strict schema, keyboard navigation, copy / paste. Like Airtable.
• 📋 Kanban, calendar, dashboard and timer views on any table.
• 🧰 Templates: issue tracker, CRM, project tasks, time tracker and more.
• 🌐 Websites: design pages in the browser, publish in one click, roll back any version.
• 🗂️ Virtual drive: mount your drive as a folder in Finder or Explorer (desktop app).
• ✨ AI with MCP support, any model via OpenRouter or local Ollama.
• 🧩 Apps: custom screens in plain JavaScript, backed by your own data.
• 🔌 Plugins and integrations: Wasm plugins, and syncing from GitHub and more.
• 💬 Group chat: channels with attachments, search and replies.
• 🎥 Meetings: video calls with shared notes and presence.
• 🎨 Canvas: an infinite drawing surface, shared live.
• 📂 Files: upload, download and preview attachments.
• 🔧 Custom data models: your own classes and properties in the Ontology Editor, shared as Atomic Schema.
• 🔒 Encrypted at rest: each agent's local database is encrypted under their own key.
• 💾 Versioning: full history, every write a signed Atomic Commit.
• 🔎 Full-text search: typeahead and fuzzy, often under 3ms, same index on server and client.
• 🔐 Authorization: read / write rights, hierarchies and invite links.
• ⚙️ RESTful API with JSON-AD, plus RDF, Turtle, N-Triples and JSON-LD.
• 📖 Pagination, sorting and filtering with Atomic Collections.
• 🚀 Fast: under 1ms median response time, powered by actix-web and redb.
• 🪶 One binary (~70MB): server, web app, search, database and automatic HTTPS.
• 💻 Runs everywhere: linux, windows, mac, arm, plus desktop and mobile apps.
• 📚 Libraries for JavaScript, React, Svelte, Rust and Dart / Flutter.

Check out the documentation for installation instructions, API docs, and more.

Issues and PRs are welcome! And join our Discord! Read more in the Contributors guide.

Atomic Data and AtomicServer have been supported by NLnet through the NGI Assure, NGI0 Entrust and NGI0 Commons funds, with financial support from the European Commission's Next Generation Internet programme, and through Eurostars. This is a large part of why the project is MIT licensed and has no proprietary core. Details and grant agreement numbers.

The code is MIT licensed — fork it, modify it, sell it.

The Atomic names and logos are not covered by that grant; they are reserved. Trademark is what lets the code stay permissively licensed, so a fork can do anything except present itself as the official Atomic Server. See TRADEMARKS.md for the policy and brand/ for the artwork, which is the single source of truth for every icon across all the Atomic apps.