Enterprise AI Capability Platform for Java/Spring Boot. Register business APIs as governed AI capabilities, compose Agents with GraphSpec, Runtime, MCP/A2A/Gateway, RunOps and Trace. This framework addresses the challenge of integrating AI agents with existing enterprise systems by transforming business APIs into manageable AI assets. It targets Java/Spring Boot developers seeking to operationalize AI securely and controllably, covering the full lifecycle from capability registration to runtime governance.

How It Works

ReachAI establishes a core loop for AI capability management. Java Spring Boot applications use annotations (@AiCapability, @AiParam) to declare business APIs as AI capabilities, registered via ai-spring-boot-starter. This starter synchronizes capability snapshots and SDK graphs to a central registry for review and cataloging. The Agent Studio offers a visual canvas to orchestrate capabilities, LLMs, and other nodes into agents. The runtime enforces governance policies (ACL, side-effect levels, preflight checks) and provides comprehensive tracing via RunOps.

Quick Start & Requirements

• Installation: Clone repo, start infrastructure (docker-compose.infra.yml), initialize DB (sql/init.sql), build backend (mvn clean install -DskipTests), run services (ai-model-service, ai-skills-service, ai-agent-service), and start admin frontend (npm install, npm run dev).
• Prerequisites: Java 17, Spring Boot 3.4, MySQL 8, Redis 7, Milvus 2.4, Nacos, Node.js/npm.
• Access: Admin UI at http://localhost:5200.
• Links: docs/ directory.

Highlighted Details

• API Registration: Annotate Java Spring Boot interfaces/methods as AI-governed capabilities.
• Agent Studio: Visual orchestration for agents, supporting LLMs, Tools, HTTP, interactive nodes, and AI-assisted workflow generation/modification.
• Runtime Governance: Enforces security via Tool ACL, Guard/Preflight checks, and comprehensive Trace/RunOps for auditing.
• Embedded Chat SDK: Securely embed AI chat experiences within business application pages, managing user context and actions.
• Enterprise Context: Manages model instances, knowledge bases, and business indexes for trusted agent execution.

Maintenance & Community

• Community: QQ Group: 1073839193.
• Development: Ongoing work includes enhanced Guard runtime, example systems, and asset marketplace features.

Licensing & Compatibility

• License: MIT License.
• Compatibility: Permissive MIT license suitable for commercial use and integration.

Limitations & Caveats

• In-Progress Features: Advanced Guard runtime features, comprehensive example systems, and a mature asset marketplace are still under active development.
• Legacy Naming: Some internal components may retain legacy naming (e.g., skill) while the product semantic is Capability.
• Setup Complexity: Requires deploying multiple infrastructure services (MySQL, Redis, Milvus, Nacos) and backend Java applications.