AgentBuilder makes it easy to build a knowledge base by uploading documents or importing online data. This knowledge base can then be integrated into your AI applications for advanced Q&A and retrieval tasks.
Steps to Create a Knowledge Base
-
Create a Knowledge Base
- Go to the Knowledge section in AgentBuilder and click Create Knowledge.
- Choose to import either local document files or online data.
-
Choose Chunking Mode
- Select a chunking mode and preview the splitting results. This step preprocesses and structures your content by dividing long texts into smaller, manageable chunks.
-
Configure Indexing and Retrieval
- Set the indexing method and retrieval settings. When a user query is received, the knowledge base searches documents using your chosen retrieval method and extracts the most relevant content chunks.
-
Wait for Embedding
- Allow time for the system to complete chunk embeddings.
-
Link to Your Application
- Once finished, link the knowledge base to your application. You can now build LLM-powered Q&A features based on your knowledge base.
Importing Text Data
- Upload Local Files
- Drag and drop or select files to upload. Batch upload limits depend on your subscription plan.
- The upload size limit for a single document is 15MB.
- Import Data from Website
- Crawl content from public web pages using third-party tools like Jina Reader and Firecrawl. Parsed content is imported as Markdown into the knowledge base.
All image links use the full domain. Steps use symbols, and the content is rewritten for clarity and conciseness.