What makes GAIA different from every other AI assistant
GAIA is the only AI assistant that's fully open source, self-hostable, and designed to manage your entire work life — without sending your data anywhere.
MIT Licensed
The entire codebase — backend, frontend, and agents — is published under the MIT license. Use it, fork it, modify it. No restrictions.
Self-Hostable via Docker
One `docker compose up` command brings up the entire stack. No complex cloud setup, no vendor lock-in. Run it on any VPS, home server, or cloud VM.
Data Stays on Your Servers
Your emails, calendar events, tasks, and conversation history are stored in databases you own and control. Zero telemetry to GAIA's infrastructure.
No Vendor Lock-In
Swap the LLM backend, add custom tools, modify agent behavior, or integrate internal systems. GAIA's architecture is designed for extensibility.
What GAIA actually does for you
A proactive AI assistant that handles the repetitive work across every tool you use — so you can focus on what matters.
- ✔Triages your inbox — labels emails by urgency, drafts replies, archives noise
- ✔Manages your calendar — schedules meetings, handles conflicts, sends invites
- ✔Tracks your tasks — syncs with Todoist, Linear, GitHub Issues, and more
- ✔Sends proactive briefings — your day's priorities at 7am, without asking
- ✔Connects 50+ tools via MCP — Slack, Notion, HubSpot, GitHub, Google Workspace
- ✔Runs automated workflows — recurring tasks, follow-ups, reports, on a schedule
Why open source matters for an AI assistant
Privacy, auditability, and true ownership. The four pillars that make self-hosted GAIA fundamentally different from closed AI tools.
Privacy
Closed AI assistants send your emails, calendar events, and work context to corporate servers for processing. With GAIA self-hosted, your data never leaves your infrastructure.
Customization
Need a custom integration with your internal tools? Want to change how the agent reasons about your tasks? You have full access to the code — modify anything.
No per-seat pricing
ChatGPT charges per message. Copilot charges per seat. Self-hosted GAIA has no usage-based charges. Pay for compute, not for AI assistance.
Auditability
You can read every line of code that processes your data. No black boxes, no hidden telemetry, no undisclosed data sharing.
GAIA (open source) vs closed AI assistants
See how GAIA stacks up against ChatGPT, Copilot, and Notion AI across the features that matter most.
| Feature | GAIA | ChatGPT | Copilot | Notion AI |
|---|---|---|---|---|
| Source code | Open source (MIT) | Closed source | Closed source | Closed source |
| Self-hostable | Yes — Docker Compose | No | No | No |
| Data location | Your servers | OpenAI's servers | Microsoft's servers | Notion's servers |
| Pricing model | Free (self-host) / cloud tier | Per-message / subscription | Per-seat subscription | Per-seat subscription |
| Email management | Yes — full triage + drafts | No | Limited (Outlook only) | No |
| Calendar management | Yes — schedule, reschedule | No | Yes (M365 only) | No |
| Proactive actions | Yes — acts without prompting | No | Limited | No |
| Integrations | 50+ via MCP | GPT plugins / limited | M365 ecosystem | Notion only |
How to self-host GAIA in 3 steps
No Kubernetes required. No cloud account needed. Just Docker. Get GAIA running on your own server in minutes.
- 1
Clone the repository
Run `git clone https://github.com/theexperiencecompany/gaia.git` to get the full source code on your machine.
- 2
Configure your environment
Copy `.env.example` to `.env` in the `apps/api` directory. Add your LLM API key, OAuth credentials for Gmail/Calendar, and any other integrations you want to enable.
- 3
Deploy with Docker Compose
Run `cd infra/docker && docker compose up` to start all services — the API, worker, databases, and message broker — in one command. GAIA is running on your server.
Frequently asked questions
Common questions about self-hosting GAIA.
MIT
open source license
1 cmd
docker compose up
50+
integrations
free
to self-host
Challenges Open Source Users Face Every Day
Closed AI assistants send your sensitive emails and work context to corporate servers with no visibility into how it's used.
Per-seat and per-message pricing makes AI assistants prohibitively expensive for teams and power users.
You can't audit, modify, or trust a black-box AI system with your most sensitive professional data.
Vendor lock-in means your workflows break the moment a service changes pricing, shuts down, or deprecates an API.
No way to integrate internal tools, custom data sources, or proprietary systems with closed platforms.
How GAIA Helps Open Source Users
MIT Licensed
The entire codebase — backend, frontend, and agents — is published under the MIT license. Use it, fork it, modify it. No restrictions.
Self-Hostable via Docker
One `docker compose up` command brings up the entire stack. No complex cloud setup, no vendor lock-in. Run it on any VPS, home server, or cloud VM.
Data Stays on Your Servers
Your emails, calendar events, tasks, and conversation history are stored in databases you own and control. Zero telemetry to GAIA's infrastructure.
No Vendor Lock-In
Swap the LLM backend, add custom tools, modify agent behavior, or integrate internal systems. GAIA's architecture is designed for extensibility.
Frequently Asked Questions
GAIA for Other Roles
AI Chief of Staff
See how GAIA acts as your proactive AI chief of staff — managing your entire workday autonomously.
Inbox Zero with AI
How GAIA automatically triages your Gmail and reaches inbox zero every day.
GAIA vs Competitors
See detailed head-to-head comparisons of GAIA against other AI and productivity tools.


