TinyAGI is an agent teams orchestrator that enables one person to run multiple teams of AI agents collaborating simultaneously across isolated workspaces.
The tool addresses the challenge of coordinating multiple specialized AI agents working together on complex tasks. It allows agents to hand off work to teammates through chain execution and fan-out patterns, with each team operating in its own isolated workspace. The system maintains persistent conversation context across restarts using SQLite-backed queues with atomic transactions and retry logic, enabling 24/7 operation as a background process or Docker container.
TinyAGI suits solo founders and small teams who need AI agents to handle multiple concurrent workflows across different channels. The tool includes a browser-based dashboard called TinyOffice for managing agents, teams, tasks, and logs, plus persistent async chat rooms per team. It supports multiple AI providers including Anthropic Claude and OpenAI, with custom provider support for any OpenAI or Anthropic-compatible endpoint. The quick start requires only Node.js and installs globally with zero configuration—it auto-creates default settings and opens the dashboard immediately. A plugin system allows extending functionality with custom message hooks and event listeners.
The project maintains active development with regular engagement in its Discord community and an explicit call for contributors. The codebase is written in TypeScript and supports macOS, Linux, and Windows via WSL2, with Docker deployment available for containerized operation.
