Developer Note: For technical setup and architecture, see docs/DEVELOPMENT.md.
Automate your workflows by talking to an AI — and let it remember, search, and handle tasks like a real assistant.
Atom is an open-source, self-hosted AI agent platform that combines visual workflow builders with intelligent LLM-based agents.
Just speak or type your request, and Atom's specialty agents will plan, verify, and execute complex workflows across your entire tech stack.
Key Difference: Unlike SaaS alternatives, Atom runs entirely on your infrastructure. Your data never leaves your environment.
- Build complex workflows using just your voice
- Natural language understanding — no proprietary syntax to learn
- Real-time feedback as Atom visualizes its reasoning
- Sales Agent: CRM pipelines, lead scoring, outreach
- Marketing Agent: Campaigns, social posting, analytics
- Engineering Agent: PR notifications, deployments, incidents
- Governance: Agents progress from "Student" to "Autonomous" as they gain trust
- Rich interactive presentations (charts, forms, markdown)
- Real-time operation visibility: See exactly what agents are doing in plain English
- Multi-view orchestration (browser, terminal, canvas)
- Smart error resolution with actionable suggestions
- Complete transparency and governance integration
- Browser automation via CDP (web scraping, form filling)
- Device control (camera, location, notifications, terminal)
- Governance-first: all actions require appropriate maturity level
- Capability Recall: Agents remember your connected services
- Unified Index: Search emails, docs, tickets, and Slack instantly
- Knowledge Graph: Understands relationships, not just keywords
- Privacy First: API keys and PII automatically encrypted
- Agents progress from 'Student' → 'Autonomous' based on performance
- Maturity-based routing: Student agents blocked from automated triggers
- AI-powered training: Personalized learning scenarios with duration estimation
- Sensitive actions require approval until confidence is high
- Every action logged, timestamped, and traceable
- Real-time supervision for learning agents
- 46+ pre-built integrations: Slack, Gmail, HubSpot, Salesforce, etc.
- Multi-platform bridge: 12+ platforms (Slack, WhatsApp, Discord, Teams, etc.)
- Use
/run,/workflow,/agentsfrom your favorite chat app
- Agents build new tools on-the-fly
- Skill Runner UI to test and execute agent skills
- Real-time streaming execution
Fastest way (Docker):
git clone https://github.com/rush86999/atom.git
cd atom
docker-compose up -dAccess at: http://localhost:3000
| Department | Scenario |
|---|---|
| Sales | New lead in HubSpot → Research company → Score lead → Slack the account executive |
| Finance | PDF invoice in Gmail → Extract data → Match against QuickBooks → Flag discrepancies |
| Support | Zendesk ticket arrives → Analyze sentiment → Route urgent issues → Draft response |
| HR | New employee in BambooHR → Provision Google account → Invite to Slack → Schedule orientation |
- Self-Hosted Only: Your data never leaves your environment
- BYOK: Bring your own OpenAI, Anthropic, Gemini, or DeepSeek keys
- Encrypted Storage: Sensitive data encrypted at-rest (Fernet)
- Audit Logs: Every agent action logged and traceable
- Human-in-the-Loop: Configurable approval policies
✅ Complete backend (FastAPI) + frontend (Next.js) + desktop app (Tauri) ✅ 46+ pre-built integrations ✅ Multi-platform communication bridge (12+ platforms) ✅ Agent governance and maturity system ✅ Memory and knowledge graph ✅ Voice interface ✅ Docker deployment
- Maturity-based routing: Prevents STUDENT agents from automated triggers
- AI training proposals: Personalized learning with duration estimation
- Real-time supervision: Monitor SUPERVISED agents with intervention controls
- Action proposals: INTERN agents require human approval before execution
- Confidence boosting: Performance-based maturity progression
- Full Documentation →
- Real-time operation tracking with plain English explanations
- Multi-view orchestration (browser, terminal, canvas)
- Smart error resolution with learning feedback
- Interactive permission/decision requests
- Auto-recording: Autonomous agents automatically record sessions for governance
- AI-powered review: Analyzes recordings to update agent confidence
- Learning loop: Successful/failed patterns feed into world model
- Confidence scoring: Approved actions increase confidence, failures decrease it
- Full audit trail for compliance
- Development Guide - Technical setup and architecture
- Student Agent Training - Maturity-based routing system
- Canvas Implementation - Canvas system details
- Agent Governance - Maturity levels and approvals
- Recording System - Recording and playback
- Review Integration - Governance & learning
We welcome contributions! See CONTRIBUTING.md for guidelines.
- Documentation:
docs/directory - Issues: GitHub Issues
- License: AGPL v3 - See LICENSE.md
Built with ActivePieces | LangChain | FastAPI | Next.js
Experience the future of self-hosted AI automation.
⭐ Star us on GitHub — it helps!
