Yao Agents is a self-hosted agent orchestration platform that lets you manage AI agents and workspaces across all your devices from a single interface.
The tool addresses the fragmentation of AI agent management by providing a unified workspace where agents run locally on your own machines while remaining accessible from desktop, mobile, browser, or API. Each workspace is isolated, allowing work to stay separate while accumulating into a knowledge base that agents can reference. Tasks created through conversation become board items that agents can act upon, improving over time as they learn the specific patterns of work. The platform integrates with large language models like DeepSeek and Claude, supporting tool use and agent skills through standard APIs with both SSE and WebSocket support.
Yao Agents suits teams and individuals who want to keep agent execution local while maintaining centralized management across multiple devices. It works well for workflows where agents need to learn task-specific patterns and where integration with existing applications is important. The open API design allows expert and task agents to be embedded into custom applications or shared with others, making it suitable for businesses building agent-powered features into their own systems.
The project shows active development with regular updates to its core agent orchestration capabilities and platform support. The codebase maintains focus on expanding device compatibility and agent integration options. Development includes ongoing work on the desktop and mobile applications alongside the backend infrastructure. The project maintains documentation and provides multiple distribution channels for different platforms.
