Exploring JoyAgent-JDGenie: The First Product‑Grade Open‑Source Multi‑Agent System

JoyAgent‑JDGenie is an open‑source, product‑grade multi‑agent system released by JD Cloud, touted as the industry’s first fully functional multi‑agent product.

It offers high completeness, lightweight, end‑to‑end capabilities, and provides the entire product stack—including front‑end, back‑end, framework, engine, and core sub‑agents—100% open source.

The project’s clear mission is to solve the “last mile” problem of quickly building multi‑agent products, allowing developers to launch an enterprise‑grade multi‑agent application in about five minutes.

Key advantages include plug‑and‑play usage, no vendor lock‑in (it can be deployed locally without binding to specific cloud providers), and support for various models such as DeepSeek.

System Architecture

The architecture follows a layered design with multi‑agent collaboration:

Model and Tools Layer (provides low‑level capabilities for agents):

LLM (Large Language Model) : e.g., deepseekV3, gpt4.1 – the core “thinking” engine.

NLP Tools : e.g., web search, browser use – enable information retrieval and interaction.

Report Tools : e.g., html tool, ppt tool – generate formatted outputs.

Memory Layer (persistent storage of experience and knowledge): stores conversation history, task history, user profiles, and domain knowledge bases.

Agent Layer : includes agent reasoning and planning, an AgentBase library, and Multi‑Agent interaction for collaborative work.

Agent Application Layer : top‑level applications such as DataAgent, Genie (general assistant), AI interview assistants, etc., covering diverse business scenarios.

The system achieves efficient task processing and tool evolution through multi‑level, multi‑mode thinking, cross‑task workflow memory, and an atomic‑tool auto‑decomposition mechanism.

Recently, JoyAgent‑JDGenie added a DataAgent capability.

Deployment can be performed via a one‑click Docker solution or manual setup, though users have reported issues such as Docker build failures and runtime problems.

Project repository: https://github.com/jd-opensource/joyagent-jdgenie