Skip to content

AtomicBot-ai/atomicbot

Β 
Β 

Folders and files

NameName
Last commit message
Last commit date

Latest commit

Β 

History

18,160 Commits
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 

Repository files navigation

Atomic Bot

Atomic Bot

Your personal AI assistant β€” right on your desktop.

Download. Install. Done. No terminal, no Homebrew, no dependencies to manage.

One app. All your AI models. All your tools. All your messengers. Batteries included.

npm version CI license

Get Atomic Bot Β· Download for macOS Β· Download for Windows

macOS Windows

Atomic Bot demo

What is Atomic Bot?

Atomic Bot is the easiest way to run a personal AI assistant. It's a native desktop app β€” not a CLI tool, not a Docker container, not a browser extension. Download the installer, open it, and you're ready to go. No terminal. No Homebrew. No npm install. Everything you need is bundled inside.

It connects the best AI models to the tools and messengers you already use β€” without switching tabs, copying text, or juggling multiple apps. Pick your AI provider, and start chatting. Atomic Bot handles the rest: it reads your messages across Telegram, Slack, Discord, and more β€” and replies for you with the power of Claude, GPT, Gemini, or any model you choose.

Why Atomic Bot?

Problem Atomic Bot
CLI tools need terminal, Homebrew, Node.js Just download and run β€” everything is bundled, zero dependencies
Switching between ChatGPT, Claude, Gemini tabs One app, all models β€” switch with a click
Copy-pasting AI answers into messengers Auto-replies across Telegram, Slack, Discord, and more
No access to your files, notes, calendar Built-in skills β€” Google Workspace, Notion, GitHub, Obsidian, Apple Notes & Reminders
Voice input requires extra tools Built-in Whisper β€” hold a button and talk, locally or via API
AI can't run commands on your machine Embedded terminal with exec approval β€” AI runs commands, you stay in control

Features

Chat with any AI model

Use the best model for the task. Switch providers and models on the fly.

  • Anthropic β€” Claude 4.6 / Opus / Sonnet / Haiku
  • OpenAI β€” GPT-5.2 etc
  • Google β€” Gemini 2.5 Pro / Flash
  • xAI β€” Grok
  • OpenRouter β€” 200+ models in one place
  • NVIDIA NIM, Venice AI, MiniMax, Moonshot, Kimi Coding, Z.ai and more

Bring your own API key or use a managed subscription β€” your choice.

Talk to your AI

Hold the microphone button and speak. Atomic Bot transcribes your voice using Whisper β€” either locally (no data leaves your machine) or via the OpenAI Whisper API for maximum accuracy. Three local model sizes: Small, Medium, Large.

Connect your messengers

Receive and reply to messages across your favorite platforms β€” all through one AI-powered interface.

  • Telegram β€” bot token setup in seconds
  • Slack β€” Socket Mode, no public endpoint needed
  • Discord Β· WhatsApp Β· Signal Β· iMessage Β· Matrix Β· Microsoft Teams β€” coming soon

Skills & integrations

Atomic Bot doesn't just chat β€” it does things. Enable skills to give your AI superpowers:

Skill What it does
Google Workspace Read & draft emails, check calendar, manage inbox
Notion Search pages, query databases, create content
GitHub Browse PRs, issues, create workflows
Obsidian Search your vaults, read and write notes
Trello Manage boards, cards, and lists
Apple Notes Create and search notes (macOS)
Apple Reminders Add and manage reminders (macOS)
Web Search Search the web with advanced queries
Media Analysis Understand images, audio, and video
ElevenLabs Generate natural speech
Image Generation Create images from text prompts
Custom Skills Upload your own skill packages

Embedded terminal

A full terminal lives inside Atomic Bot. Your AI can suggest and run shell commands β€” with your explicit approval every time. Allow once, always allow, or deny. You stay in control.

Smart exec approval

When the AI wants to run a command, you see exactly what it plans to do. One click to approve, deny, or auto-approve that command pattern. Security by default, convenience by choice.

Auto-updates

Atomic Bot keeps itself up to date. New version available? A banner appears, you click download, and restart when ready. Release notes included.

Backup & restore

Export your entire configuration as a ZIP snapshot. Restore it on a new machine or after a reset β€” all your settings, keys, and skills come back instantly.

Getting started

No terminal required. No package managers. No config files. Just three steps:

  1. Download Atomic Bot from atomicbot.ai β€” one file, nothing else to install
  2. Open β€” drag to Applications (macOS) or run the installer (Windows). All runtimes are bundled
  3. Pick your AI and go β€” choose a provider, enter your API key, start chatting

Want to connect messengers or enable skills? The app walks you through it with a built-in setup wizard.

Platforms

Platform Architecture Format
macOS Apple Silicon (arm64) .dmg
Windows x64 .exe (NSIS installer)

Built with

Component Technology
Shell Electron
UI React, Redux Toolkit, React Router
Build Vite, TypeScript, esbuild
Terminal xterm.js, node-pty
Voice Whisper (local + API)
Markdown react-markdown, KaTeX (math rendering)
Updates electron-updater

Links

Open source foundation

Atomic Bot is built on top of OpenClaw β€” an open-source personal AI assistant platform licensed under MIT. OpenClaw provides the gateway, agent runtime, channel integrations, and skills engine that power Atomic Bot under the hood.

License

Packages

 
 
 

Contributors

Languages

  • TypeScript 86.9%
  • Swift 8.0%
  • Kotlin 1.9%
  • JavaScript 1.0%
  • Shell 1.0%
  • CSS 0.8%
  • Other 0.4%