langchain4j-aideepin

LangChain4j-AIDeepin is a backend service that provides a comprehensive AI-assisted workflow platform with multi-session interactions, image generation, knowledge retrieval (RAG with vector search and graph search), AI-powered web search workflows, and speech capabilities (ASR & TTS). The MCP server arrangement shown here runs the service in Docker, exposing the API gateway on port 8888 of the host (mapped to port 9999 inside the container). Once started, you can access the server APIs at http://localhost:8888. Use the platform through its management interfaces and the various endpoints exposed by the backend to manage model providers, perform vector or graph searches, and orchestrate AI workflows. The system integrates multiple model providers and enables configuring models, embeddings, and search indexes, enabling seamless AI-assisted collaboration and knowledge work.

Prerequisites:

• Docker installed on your host
• Basic familiarity with containerized services

Installation steps:

1. Ensure Docker is running on your machine.

2. Pull the image (as defined in the README execution):

   docker build . -t aideepin:0.0.1

3. Run the container with appropriate environment and volume mappings:

   docker run -d --name=aideepin -p 8888:9999 -e APP_PROFILE=[dev|prod] -v "/data/aideepin/logs:/data/logs" aideepin:0.0.1

4. Verify the service is healthy by checking logs or hitting the API endpoint:

   curl -sS http://localhost:8888/health || echo "check logs via docker logs aideepin"

Notes:

• If you prefer building locally, you can clone the repository, build with Maven as described in the project, package the jar, and run the jar directly. The README provides jar run commands for non-Docker deployment.

Tips and common issues:

• Ensure the host directory /data/aideepin/logs exists or is writable by Docker to avoid log-write failures.
• The APP_PROFILE env var controls the active Spring profile; set it to dev or prod as appropriate.
• If you modify configuration, restart the container to pick up changes.
• When using the docker approach, ensure port 8888 on the host is available and not blocked by a firewall.
• For resource-heavy operations (e.g., large language models, vector databases, or image generation), allocate sufficient CPU/memory to the container and adjust Java memory usage if running outside Docker.
• The deployment supports integrating multiple model platforms and has a guided setup in the project wiki; refer to the detailed steps there for configuring model providers, vector databases, and search indexes.