Knowledge Base

The Knowledge Base is the collection of all data you provide to your chatbot. The quality and comprehensiveness of this data directly determine the accuracy and usefulness of your chatbot's answers.

Managing Data Sources

The main view of this page is a table listing all your data sources. Each source has a status:

• Pending: The source has been submitted and is waiting to be processed.
• Training: The AI is actively reading, chunking, and indexing the content.
• Ready: The source has been successfully trained and is now being used by the chatbot to answer questions.
• Error: Something went wrong during the training process. This could be due to an inaccessible URL, a corrupted file, or content that exceeds your plan's limits.

Source Types

Blynk supports several types of data sources:

• Files: Upload documents like PDFs, DOCX, and TXT files. Ideal for internal documentation, product manuals, and policy documents.
• Website URL: Provide a starting URL, and our scraper will crawl that page and any linked sub-pages on the same domain to gather content. Perfect for help centers and FAQ pages.
• Q&A: Manually input specific question-and-answer pairs. This gives you precise control over the chatbot's response to common or important questions.