ageerle/ruoyi-ai

RuoYi AI – Enterprise‑grade AI Assistant Platform

What it is

• A full‑stack, open‑source platform that lets you build and run AI assistants with multi‑agent collaboration, RAG (retrieval‑augmented generation), and visual workflow orchestration.
• Built on Spring Boot 3 (backend) and Vue 3 (frontend) and integrates the Java LLM framework Langchain4j.

Key capabilities

Area	What you get
Model Management	Connect to many LLM providers (OpenAI, DeepSeek, Zhipu, MIMO, Baidu Bailian) and multimodal models; also plug‑in external services like Coze, DIFY, FastGPT, RAGFlow.
Knowledge Management	Local RAG pipeline with vector stores (Milvus, Weaviate, Qdrant) and built‑in document parsers for PDF, Word, Excel, images.
Tool Management	MCP protocol for tool calls, extensible “skills” ecosystem.
Workflow Orchestration	Drag‑and‑drop visual designer; nodes can call models, send email, pause for manual review, etc.; results streamed via SSE.
Multi‑Agent	Langchain4j‑based agent framework with a Supervisor mode that can route tasks to different decision models and combine tools dynamically.

Technical stack

• Backend: Spring Boot 3.5.8 + Langchain4j
• Frontend: Vue 3 + Vben Admin + Element‑Plus‑X
• Data: MySQL 8, Redis, vector DBs (Milvus/Weaviate/Qdrant), MinIO for object storage
• Security: Sa‑Token + JWT
• Realtime: WebSocket + Server‑Sent Events

How to run

• One‑click Docker – docker compose -f docker-compose-all.yaml up -d starts the backend, admin UI, user UI, and all dependencies (MySQL, Redis, Weaviate, MinIO). Ports are pre‑mapped (e.g., admin UI 25666, user UI 25137, API 26039).
• Step‑by‑step – Build each service from source with separate docker‑compose files if you need custom code.

Main repositories

• ruoyi-ai – backend services
• ruoyi-web – user‑facing frontend
• ruoyi-admin – admin panel
• ruoyi-drama, ruoyi-copilot, ruoyi-uniapp – demo apps and mini‑programs

Typical use cases

• Enterprise internal assistants that retrieve company knowledge via RAG.
• Multi‑agent task automation (e.g., coding help, short‑drama generation, email drafting).
• Rapid prototyping of AI‑augmented workflows with a visual designer.

Getting started

1. Clone the repo and check out the v3.1.0 tag.
2. Copy the example .env file, set RUIYI_VERSION=v3.1.0.
3. Run the Docker compose command above.
4. Open the admin UI (http://<host>:25666, default credentials admin/admin123) to configure model API keys and vector stores.
5. Explore the live demos (drama, user frontend, coding) for ready‑made examples.

Documentation & community

• Full docs: https://doc.ruoyiai.chat/
• Live demos: https://web.ruoyiai.chat/ (QR‑code login)
• Community chat via WeChat, Douyin, QQ (QR codes in README).

License

• MIT License.

Why it matters RuoYi AI bundles the pieces most projects have to assemble separately—LLM access, vector search, tool calling, and a visual workflow UI—into a single, deployable stack. This makes it practical for enterprises to roll out AI assistants without writing a lot of glue code.