langchain4j-aideepin

LangChain4j-AIDeepin (得应AI) is a self-deployed AI productivity platform. The README says it can support teams in technical R&D, product design, HR, finance and IT questions, system or product consulting, and customer service scripts. It has three parts: this backend repository, a user web app and an admin web app, with docs on the wiki and an online demo site.

Features listed include multi-session chat with multiple roles, image generation (text to image, retouching, image to image), knowledge-base RAG using vector search or graph search, LLM-based web search, an AI workflow builder, an MCP service marketplace, long-term memory, and speech: ASR and TTS with selectable question and reply formats and voices. Files can be stored locally or in Alibaba Cloud OSS. A table shows which functions each model platform supports; the platforms named are Lingji (DashScope), OpenAI, SiliconFlow, Ollama, DeepSeek and Qianfan.

The backend uses JDK 17, Spring Boot 3.0.5, langchain4j and langgraph4j, with PostgreSQL (pgvector and Apache AGE extensions) or Neo4j 5.26.4+ for vectors and graphs. The front ends use Vue 3, Vite, TypeScript, Pinia and Naive UI. Deployment means creating the database, running `docs/create.sql`, enabling at least one model platform by setting its API key, then building, with a Docker build also described. The licence is MIT and the last push was 2026-10-09. The documentation is in Chinese with an English README as well. This review is based on the README and repository data; we have not run it ourselves.