An open-source Go-based application engine integrating autonomous agent framework, GraphRAG knowledge retrieval, and full-stack runtime in a single binary for building web apps, REST APIs, and edge intelligence services.
Yao is a high-performance application engine designed for the AI-native era. It integrates an Agent framework, GraphRAG, native MCP support, web server, and ORM into a single Go binary, allowing developers to build everything from admin panels to intelligent agents without complex environment setup.
Core Capabilities#
Autonomous Agent Framework#
- Six-Stage Execution Model: Inspiration → Goals → Tasks → Run → Deliver → Learn
- Three Trigger Modes: Clock (scheduled), Human (email/message), Event (webhook/database)
- Continuous Learning: Iterative optimization with private knowledge bases
- Multi-Agent Collaboration: Complex task coordination across multiple agents
Built-in GraphRAG#
- Vector retrieval and knowledge graph entity relationship search
- Hybrid retrieval mode (vector + graph)
- Ideal for private knowledge base construction and intelligent Q&A
Native MCP Support#
- Map Yao processes as MCP tools
- Connect external servers via SSE/STDIO
- Declarative input/output schemas
Full-Stack Runtime#
- All-in-One executable (Data, API, Agent, UI)
- Built-in V8 engine supporting TypeScript/JavaScript
- Single binary deployment, no Node.js, Python, or container dependencies
- ARM64/x64 edge device support
DSL & Generative Programming#
- Human-readable, AI-friendly DSL
- AI-generated, visually edited, and hand-written code coexist in the same project
Built-in Infrastructure#
| Component | Function |
|---|---|
| HTTP Server | Built-in web service |
| Process Engine | Workflow scheduling engine |
| Database ORM | Multi-database support |
| Template Engine | SUI template rendering |
| AI Connector | OpenAI and other model integrations |
| Store KV | Key-value storage |
| Schedule | Scheduled tasks |
| WebSocket | Real-time communication |
Installation#
Linux/macOS:
curl -fsSL https://yaoapps.com/install.sh | bash
Windows (WSL2 required):
powershell -c "irm yaoapps.com/install.ps1 | iex"
Docker:
docker pull yaoapp/yao:latest # Production image (Alpine)
docker pull yaoapp/yao:latest-dev # Dev image (Ubuntu 22.04)
Quick Start#
mkdir /path/to/your/project
cd /path/to/your/project
yao start
Default access URL: http://127.0.0.1:5099
Common CLI Commands#
yao version --all
yao run utils.app.Ping
yao run models.tests.pet.Find 1 '::{}'
yao run scripts.tests.Hello 'Hello, Yao!'
Use Cases#
- IoT applications requiring offline operation or ARM64/x64 edge deployment
- Internal tools or SaaS prototypes needing rapid Admin Panel and REST API generation
- Multi-agent collaboration systems with private knowledge base continuous learning
- Event-driven intelligent agents and automation bots
Important Notes#
- Licensed under Modified Apache License 2.0 with trademark and authorization verification clauses
- Commercial use may require additional authorization; compliance evaluation recommended
- Primary documentation in English; Chinese README available