moyangzhan/langchain4j-aideepin

What it solves

LangChain4j-AIDeepin is an AI application platform designed to help developers rapidly build intelligent business assistants. It provides a unified interface to integrate various AI capabilities—such as chat, knowledge retrieval, and complex workflows—without needing to build these components from scratch.

How it works

The platform is built using a Spring Boot backend powered by LangChain4j and LangGraph4j. It consists of three main components: a server for core logic, an admin dashboard for management, and a user-facing web application. It integrates with multiple model providers (OpenAI, Qwen, SiliconFlow, Ollama, and DeepSeek) to handle text, image, and audio tasks.

Who it’s for

It is intended for developers and businesses looking to deploy AI assistants that require a combination of RAG, memory, and structured workflows.

Highlights

• AI Workflow Editor: A visual editor supporting conditional branching and parallel execution for LLM calls and human feedback.
• Knowledge Base (RAG): Combines vector search and knowledge graph retrieval for better information accuracy.
• MCP Service Marketplace: Allows the integration of Model Context Protocol (MCP) services to extend AI capabilities with external tools.
• ** uma Memory Management**: Implements both short-term and long-term memory to personalize responses based on historical context.
• Multimodal Support: Integrates ASR (Automatic Speech Recognition) and TTS (Text-to-Speech) for flexible voice and text interactions.