Sovereign Infrastructure for Autonomous AI
Deploy, orchestrate, and govern AI agents that plan, reason, and execute independently—with persistent memory and hierarchical delegation.
Built for Deep Agent Systems
Hierarchical Delegation
Orchestrator-worker patterns with context isolation. Agents can spawn and coordinate sub-agents for complex multi-step tasks.
Persistent Memory
Virtual filesystems, vector databases, and episodic storage. Your agents remember everything across sessions.
Multi-Runtime Support
Run agents on Claude Code or Gemini CLI. Isolated Docker containers per agent with dedicated resources and 16-tool MCP integration.
Autonomous Operations
Cron-based scheduling, real-time monitoring, self-healing capabilities, and OpenTelemetry metrics for cost and productivity tracking.
Case studies
See the impact of AI system implementation.
Frequently Asked Questions
What makes Trinity different from other agent frameworks?
Trinity focuses on 'System 2' AI—agents that plan and reason autonomously rather than just react. It provides persistent memory, hierarchical delegation, and production-grade infrastructure for agents that operate independently over time.
What are the system requirements?
You need Docker & Docker Compose v2+, plus an Anthropic API key (for Claude) or Google API key (for Gemini). Installation is a single command.
Is Trinity free to use?
Trinity is free for personal use, research, education, and nonprofits under the Polyform Noncommercial License. Commercial use requires a separate license—contact us for details.
Can agents communicate with each other?
Yes. Trinity supports agent-to-agent communication with permission controls, shared folders for state synchronization, and parallel task execution for orchestrator-worker patterns.
Ready to Deploy Autonomous Agents?
Get started with Trinity in minutes. One-line install, web-based setup wizard.
View on GitHub