AIDeepin is an AI-powered productivity suite designed for enterprise and team use, offering features like multi-session chat, image generation, RAG-based knowledge bases, and workflow automation. It targets businesses and teams looking to enhance efficiency in areas such as technical R&D, product design, and customer support.

How It Works

The backend is built with Java 17 and Spring Boot, leveraging langchain4j for AI orchestration. It integrates with various LLMs (DeepSeek, OpenAI, Qwen, Ernie, Ollama) and image models (DALL-E, Tongyi Wanxiang). Data storage utilizes PostgreSQL with pgvector for vector embeddings and Apache AGE or Neo4j for graph data, offering flexibility in data management.

Quick Start & Requirements

• Installation: Build and run the backend JAR or use Docker.
• Prerequisites: Java 17, Maven, PostgreSQL, and optionally Neo4j, Redis, and an email service.
• Configuration: Requires setting up AI model API keys and potentially search engine configurations via SQL updates or the admin UI.
• Documentation: 中文文档 | English

Highlighted Details

• Supports RAG with both vector and knowledge graph databases (PostgreSQL+pgvector, Apache AGE, Neo4j).
• Includes multi-modal capabilities with image generation and editing features.
• Offers a workflow engine for custom AI agents and task automation.
• Integrates with multiple LLM providers and open-source models via Ollama.

Maintenance & Community

The project is maintained in developers' spare time. Community support is available via GitHub Issues.

Licensing & Compatibility

The project appears to be under a permissive license, but specific details are not explicitly stated in the README. Compatibility for commercial use or closed-source linking would require verification of the exact license.

Limitations & Caveats

The project is primarily developed in spare time, which may impact response times for support. Some features like Bing and Baidu search integration are marked as TODO. Advanced RAG features like query compression and re-ranking are also listed as future work.