What is the Yanxi Agent Platform
The Yanxi Agent Platform is a one-stopAI agent development platformlaunched by JD. Regardless of whether users have a programming background, they can quickly build AI model-based agents to handle everything from Q&A to complex business logic. The platform integrates multiple large models, provides an algorithm library and tools, and supports the rapid implementation of industry applications. Currently, more than 3,300 agents are active within JD, and more than 100 industry solution templates have been accumulated.
Main features of the Yanxi Agent Platform
-
Accesslarge models: The platform has integrated dozens of large models and supports users in selecting different models according to business needs.
-
Low-cost rapid building: Regardless of whether users have a programming background, they can quickly build AI model-based agents.
-
Industry solution templates: The platform has accumulated more than 100 industry solution templates, supporting the rapid implementation of industry applications.
-
Algorithm library and tool library: Through plugin capabilities, the platform provides thousands of algorithms and tool capabilities, such as data analysis, NL2SQL, etc.
How to use the Yanxi Agent Platform
-
Registration and login: Users need to visit the official website of the JD Cloud Yanxi Agent Platform, register an account and log in (yanxi.jd）。
-
Selectagenttemplate: The platform provides a variety of industry solution templates, and users can choose a suitable template as a starting point according to their own business needs.
-
Configureagent: Users can configure the selected agent template, including but not limited to setting the agent's name, functions, interaction logic, etc.
-
Accesslarge models: According to business needs, users can select and access different large models on the platform, such as the Yanxi large model, GPT, etc.
-
Knowledge base access: Using Advance RAG technology, users can simply configure access to structured and unstructured data to enhance the agent's knowledge base.
-
Application of algorithm library and tool library: Users can use the algorithm and tool capabilities provided by the platform in agents, such as data analysis, NL2SQL, etc.
-
Workfloworchestration: Through workflows, the agent's plugins and large model capabilities are orchestrated and combined to guide the agent to act according to established ideas.
-
Intelligent data analysis: Using the platform's data analysis capabilities, users can query and analyze business data through natural language.
-
Testing and optimization: After the agent is built, users need to test it and optimize and adjust the agent based on the test results.
-
Deployment and application: After testing is correct, users can deploy the agent to actual business scenarios and begin using the agent to handle business problems.
