🚀 Personal AI agents, under your control.
9Lives is a local-first runtime for AI agents that lets you run autonomous agents on your own infrastructure.
Run your agents on:
- Raspberry Pi
- Personal computers
- Private servers
- Home labs
Create Lives (agents), give them Skills (tools), and orchestrate them with Crews.
User → Telegram
↓
9Lives
↓
AI agent executes workflow
↓
Email / task / report
Example use case:
Morning briefing agent that gathers news and sends you a daily report.
Most AI agent frameworks today are:
- cloud dependent
- difficult to control
- unpredictable in production
9Lives focuses on:
- local-first execution
- controllable agents
- deterministic workflows
- multi-agent orchestration
A Live is an AI agent with its own identity, memory and capabilities.
Skills extend what a Live can do.
They are simple TypeScript functions.
A Crew is a team of Lives working together.
A Lane defines how work is executed.
Each Live maintains its own contextual memory.
Define safety boundaries for agent execution.
git clone https://github.com/yourusername/9lives
cd 9lives
docker compose upOpen:
Recommended:
- Raspberry Pi 5
- 8GB RAM
- Docker
User Interfaces
↓
Lane Queue
↓
Crew Orchestrator
↓
Lives
↓
Skills
Explore the full technical documentation:
- 📖 Technical Guide — complete overview of architecture, concepts, and setup
- core runtime
- Lives
- Skills
- Crews
- Lanes
- scheduler
- dynamic skill loading
- plugin system
- skill registry
- monitoring
Contributions are welcome!
Copyright (c) 2026 Flavio Cerato
Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with the License.
Run your own AI agents. Locally. Under your control.
