Overview
Features
Large model invocation and prompt customization
Call large language models and customize prompts to fit specific use cases and model parameters.
Capability components invocation
Access 40+ Baidu ecosystem components to empower Agent applications and enhance functionality.
AI-native applications access & management
Manage AI-native apps published in AppBuilder and register local functions to link cloud components.
KnowledgeBase and RAG orchestration
Manage KnowledgeBase, perform document parsing, chunking, embedding, indexing, and retrieval for industrial-grade RAG workflows.
Workflow orchestration with multi-level abstractions
Use Message, Component, and AgentRuntime to orchestrate complex pipelines; integrate with LangChain/OpenAI ecosystems.
Monitoring and debugging tools
Provide visual tracing and detailed debug logs to assist development and production troubleshooting.
Deployment options
Deploy AgentRuntime as Flask/gunicorn API or as a Chainlit frontend; publish public APIs with appbuilder_bce_deploy.
End-to-end RAG cookbooks and examples
Access cookbooks and examples to build and customize end-to-end RAG applications.
Who Is This For?
- AI-native application developers:Build, customize, and deploy AI-powered agents and apps using the SDK's invocation, orchestration, and deployment capabilities.
- RAG/Knowledge-base engineers:Construct knowledge-base driven workflows with parsing, embedding, indexing, and retrieval for industrial-grade applications.
- DevOps/Platform engineers:Monitor production apps with tracing and debug logs; deploy and publish public APIs via Baidu Cloud.




