Yao is an open-source engine for building autonomous agents, designed to move beyond passive AI assistants. It enables the creation of proactive, self-scheduling agents that operate autonomously, functioning more like team members than simple tools. This framework allows developers to build agents triggered by events, schedules, or human input, significantly enhancing automation capabilities.

How It Works

Yao provides an event-driven, proactive autonomous agent framework, fundamentally differing from passive AI assistants. Agents are conceptualized as team members, not mere tools, with flexible entry points like email, events, and schedules, moving beyond traditional chat interfaces. The core execution follows a six-phase lifecycle: Inspiration, Goals, Tasks, Run, Deliver, and Learn, enabling a structured approach to agent operation. Dynamic multi-agent orchestration facilitates complex workflows through delegation, collaboration, and dynamic composition. Continuous learning is integrated via private knowledge bases, allowing agents to accumulate and leverage experience over time.

Quick Start & Requirements

• Install/Run: Single binary executable.
• Prerequisites: No Node.js, Python, or containers required. Built-in V8 engine for TypeScript. Runs on ARM64/x64 devices.
• Links:
  • Homepage: https://yaoapps.com
  • Quick Start: https://yaoapps.com/docs/documentation/en-us/getting-started
  • Documentation: https://yaoapps.com/docs

Highlighted Details

• Autonomous Agent Framework: Enables event-driven, proactive, and self-scheduling agents that operate autonomously.
• Flexible Triggers: Supports three distinct modes: Clock (scheduled tasks), Human (email/message interactions), and Event (webhooks/database changes).
• Six-Phase Execution Cycle: A structured workflow from initial Inspiration and Goal setting through Task execution, Delivery, and iterative Learning.
• Multi-Agent Orchestration: Agents can dynamically delegate tasks, collaborate with peers, and compose functionalities for complex problem-solving.
• Continuous Learning: Agents build private knowledge bases to accumulate experience, improving performance and adaptability over time.
• Native MCP Support: Streamlines tool integration by mapping Yao processes directly to external tools without custom adapters, supporting Process Transport, SSE, or STDIO.
• Built-in GraphRAG: Integrates advanced retrieval mechanisms, including embeddings (OpenAI/FastEmbed), knowledge graph traversal, and hybrid search for contextually rich information retrieval.
• Full-Stack Runtime: Consolidates data management, API serving, agent logic, and UI rendering within a single, self-contained executable.
• TypeScript Support: Features a built-in V8 engine, allowing agents to be developed using TypeScript.
• Edge-Ready Deployment: Designed for efficient operation on ARM64/x64 devices, facilitating deployment in diverse environments.

Maintenance & Community

No information provided in the README snippet.

Licensing & Compatibility

No information provided in the README snippet.

Limitations & Caveats

No information provided in the README snippet.