JoyAgent-JDGenie is an open-source, end-to-end, product-grade multi-agent system designed to simplify the creation of sophisticated AI agent applications. It targets developers and researchers seeking a ready-to-use solution for complex tasks, offering direct query resolution and report generation capabilities, unlike SDK-focused alternatives.

How It Works

The system employs a multi-agent framework with pluggable sub-agents and tools, supporting various design patterns like React and Plan-and-Executor. It features a high-concurrency DAG execution engine for efficiency and innovative techniques such as multi-level, multi-pattern thinking, cross-task memory, and tool evolution via auto-disassembly and reassembly of atomic tools. This approach allows for dynamic tool creation and adaptation without manual intervention.

Quick Start & Requirements

• Docker: Clone the repository, configure application.yml and .env_template with API keys and model details (e.g., DeepSeek), build the Docker image (docker build -t genie:latest .), and run the container (docker run -d -p 3000:3000 -p 8080:8080 -p 1601:1601 --name genie-app genie:latest). Access via localhost:3000.
• Manual Setup: Requires JDK 17, Python 3.11. Install dependencies using pip install uv, cd genie-tool, uv sync, and activate the environment (source .venv/bin/activate). Deployment can be done via a step-by-step guide or a one-click script (sh check_dep_port.sh, sh Genie_start.sh).
• Dependencies: LLM API keys (e.g., DeepSeek, OpenAI), Serper API key for search.
• Resources: Detailed deployment videos are provided.

Highlighted Details

• Achieves 75.15% accuracy on the GAIA Validation set and 65.12% on the Test set, outperforming notable competitors like OWL and AutoAgent.
• Offers a lightweight, self-contained solution, avoiding dependencies on proprietary cloud platforms like Alibaba Cloud Baishan or Volcano Engine.
• Provides a full product stack including frontend, backend, framework, engine, and core sub-agents (report, code, PPT, file).
• Supports flexible tool integration and custom sub-agent development by implementing the BaseTool interface.

Maintenance & Community

The project is contributed to by a team from JD CHO Enterprise Informationization Team. Contributions are welcomed via Pull Requests after signing a contributor agreement. Academic citation details are provided.

Licensing & Compatibility

The repository does not explicitly state a license in the README. Compatibility for commercial use or closed-source linking is not specified.

Limitations & Caveats

The license is not clearly defined, which may impact commercial adoption. The README mentions configuration steps for specific LLMs like DeepSeek, implying potential complexity in adapting to other models or environments.