How JD’s Open‑Source Large‑Model Tools Are Shaping the Future of Enterprise AI

With the rapid development of large‑model technology, AI tools have become essential productivity assistants that boost efficiency, solve business pain points, and extend knowledge through intelligent collaboration. Mastering these tools is now a key competitive advantage for modern professionals.

Industry Status of Large‑Model Tools

Since the emergence of ChatGPT, large models have made breakthrough progress and are widely used in text generation, intelligent customer service, and more. Leading companies follow five design principles:

Low‑code UI : Lower the usage barrier for non‑technical users.

Security (local deployment) : Private deployment prevents data leakage.

Usability (multi‑device compatibility) : One code works on many devices, reducing adaptation costs.

Scalability : Align with open‑source trends for continuous evolution.

Cost‑effectiveness : Ensure commercial sustainability.

JD’s Large‑Model Capabilities and Applications

JD has upgraded its Yanxi large‑model brand to JoyAI , offering models from 3B to 750B parameters across language, speech, and vision modalities. Enhanced inference efficiency and reduced training costs enable deployment in retail, logistics, health, and industry, serving over a million merchants. JD also built the enterprise AI Agent platform JoyAgent , the intelligent coding assistant JoyCode , and the multi‑agent collaboration framework OxyGent .

JoyAgent

JoyAgent is the industry’s first 100% open‑source enterprise‑grade intelligent agent, covering front‑end, back‑end, framework, engine, and core sub‑agents. It can be deployed locally and enables rapid AI adoption through low‑code workflow building and data‑closed‑loop optimization.

OxyGent

OxyGent abstracts tools, models, and agents as plug‑in atomic operators, allowing developers to assemble multi‑agent systems like building blocks, offering extreme extensibility and end‑to‑end decision traceability.

JoyCode

JoyCode has evolved from AI‑assisted code completion to a multi‑agent autonomous programming platform. Two success cases are highlighted: (1) Front‑end H5 development using Figma MCP + JoyCode; (2) Custom agent generation for SDK integration. Challenges such as goal drift are addressed with a “plan‑first‑execute‑later” strategy and customizable agents.

Practical Business Scenarios

Office : AI‑powered assistants like Max automate meeting minutes and task creation, moving toward a “digital employee” ecosystem.

Code Review : JD’s AI review system, built on Yanxi, DeepSeek, and AutoBots, automates line‑level checks, reduces Java defect rates by 32%, and integrates with the corporate coding platform.

Local Life : Applications include cross‑language code translation, universal knowledge labeling, and multi‑agent simulation for delivery logistics.

Logistics : Deployments such as digital‑warehouse AR glasses, intelligent outbound call systems, building‑level address parsing, and elevator‑recognition models improve efficiency and reduce error rates.

Policy Outlook

Regulatory frameworks like China’s “Interim Measures for Generative AI Services” and the EU AI Act are shaping data quality, content safety, and copyright standards. Companies must align with these policies to mitigate risk and influence standards.

Future of AI Agents

Agents will evolve toward multi‑agent collaboration, embodied digital humans, and collective intelligence, enabling complex task handling, physical interaction, and continuous self‑evolution.

Path to Embodied Intelligence

Key enablers are multimodal perception fusion, data self‑closed loops, and lightweight deployment (e.g., INT4 quantization) to run 7B models on edge chips, achieving end‑to‑end perception‑decision‑control pipelines.