ruoyi-ai and ZenoAgent
These are competitor AI agent development frameworks, with project A offering broader enterprise-grade features like multi-vendor model management and visual orchestration, while project B provides a more focused Spring Boot and Vue 3 platform with LangChain4j integration for core AI agent capabilities.
About ruoyi-ai
ageerle/ruoyi-ai
面向企业级市场的一站式AI应用开发框架，支持多厂商大模型统一接入与管理，具备安全可控的企业知识库与高精度检索优化能力，提供可视化流程编排、自主决策智能体与多智能体协同调度，兼容主流 Agent Skill 协议，同时支持微信生态扩展，帮助企业与开发者零门槛快速构建安全、高效、可落地的AI智能体应用与行业解决方案。
Built on Spring Boot 4.0, Spring AI 2.0, and Langchain4j, the platform integrates multiple vector databases (Milvus, Weaviate, Qdrant) for RAG-based knowledge retrieval and supports MCP protocol tools alongside custom Skills. Multi-agent coordination uses a Supervisor pattern with multiple decision models, while SSE streaming and WebSocket enable real-time workflow execution with document parsing (PDF, Word, Excel) and image analysis capabilities.
About ZenoAgent
Johnnyjin-haolin/ZenoAgent
Zeno Agent 是一个基于 Spring Boot 和 Vue 3 构建的 AI Agent 平台，集成了 LangChain4j 框架，提供智能对话、知识检索增强（RAG）、工具调用（MCP）等核心能力。项目采用前后端分离架构，支持流式响应、多会话管理、知识库管理等完整功能。
Implements dual-scope MCP architecture with GLOBAL (server-side) and PERSONAL (browser-based) modes, enabling hybrid tool execution while protecting user credentials; enables handling of dozens to hundreds of tools through progressive loading of tool schemas and skill summaries, bypassing LLM context window limits. Distributes human-in-the-loop confirmations across multi-node deployments using Redisson's blocking queue, with real-time SSE event streaming throughout the reasoning lifecycle including tool calls, RAG retrieval, and PERSONAL MCP task delegation.
Related comparisons
Scores updated daily from GitHub, PyPI, and npm data. How scores work