Large language models cannot directly answer questions in private knowledge domains. However, you can Large language models (LLM) cannot directly answer questions in private knowledge domains. However, you can create an agent application in Model Studio and integrate private knowledge documents to build a Q&A application that can answer questions in private domains without writing any code.
Only users who created Model Studio applications before April 21, 2025, can access the Application tab and use all its features, including applications (Agent applications, Workflow applications, Agent orchestration applications), components (Prompt, Plug-in), data (Knowledge base, Application data) features and related APIs. This feature is in preview. Use with caution in production environments.
Performance showcase
Application without a knowledge base Without a private knowledge base, the LLM cannot accurately answer questions about "Bailian phones". | Application with a knowledge base With a private knowledge base, the LLM can provide accurate answers to questions about "Bailian phones". |
Step 1: Build your first agent application (About 1 minute)
| |
| |
| |
| |
| |
|
Step 2: Build a knowledge base (About 3 minutes)
Upload knowledge documents
| |
| |
|
Create a knowledge base
| |
| |
| |
| |
|
Step 3: Add the knowledge base to your application and publish it (About 1 minute)
| |
| |
| |
|
Step 4: Test the application (About 3 minutes)
Model Studio provides a test web page for your application, allowing you to share and test it on a small scale.
| |
| |
|
What to do next
To learn about prompt writing, plug-ins, publish channels, and other application features, see Agent application.
To integrate more external tools or have the application automatically complete complex tasks and business processes, see Workflow application or Agent orchestration application. For feature comparisons and application scenarios, see Application overview.
To call Q&A applications or other Model Studio applications through API, see Call an application.
To develop highly customized RAG applications with complex interaction logic using full code, see Assistant API.