Skip to main content

Knowledge Bank Node

The Knowledge Bank node allows you to upload reference documents that can be used to provide knowledge-based contextual responses in your AI workflows. This node acts as a document repository that the AI can reference when generating responses, making it ideal for creating chatbots, tutoring systems, or any application that needs to draw from specific source materials.

Knowledge Bank node

Basic Usage​

Reference document for Knowledge Bot, providing knowledge-based contextual responses.

Inputs​

This node does not require any upstream inputs. It serves as a standalone document repository that can be referenced by downstream AI nodes.

Outputs​

  • Knowledge Output: The uploaded document(s) and their content, which can be connected to AI nodes (such as Knowledge Bot) to provide contextual information for generating responses.

Usage​

Uploading Documents​

  1. Click on the Knowledge Bank node in your flow.
  2. In the node configuration panel, click Browse to select a document from your device.
  3. Upload the document that contains the reference information you want the AI to use.
  4. The document will be processed and stored in the Knowledge Bank.

Connecting to AI Nodes​

Typical workflow:

  1. Knowledge Bank → Knowledge Bot: Connect the Knowledge Output to the Knowledge Bot's Knowledge Reference input.
  2. The Knowledge Bot will use the uploaded documents to provide informed, contextual responses based on the content.

Example Workflow​

Scenario: Creating a course-specific Q&A chatbot

  1. Knowledge Bank: Upload course materials (e.g., course-syllabus.pdf, lecture-notes.pdf).
  2. Knowledge Bank (Knowledge Output) → Knowledge Bot (Knowledge Reference): Connect the knowledge output to provide context.
  3. Knowledge Bot: Configure the bot to answer student questions using the uploaded course materials.
  4. Knowledge Bot → Additional System Prompt: Add specific instructions for how the bot should respond (e.g., "Answer questions based on the course materials. If information is not in the materials, say so.").

Result: Students can ask questions about the course, and the bot will provide answers based on the uploaded reference documents.

Preview​

When you upload a document, you can preview how the Knowledge Bank will be used in your flow:

Knowledge Bank Preview

The preview shows:

  • Knowledge Bot: The AI node that will use the knowledge
  • Knowledge Reference: Where the uploaded documents are connected
  • Bot Introduction: First message shown to users
  • Additional System Prompt: Custom instructions for the AI
  • Overwrite System Prompt: Option to completely replace default AI behavior
  • Plus User Message: Additional context from user input
  • Placeholder: Text shown in the input field
  • Temperature: Controls randomness in AI responses (lower = more focused, higher = more creative)

Supported File Types​

The Knowledge Bank supports various document formats, including:

  • PDF files (.pdf)
  • Word documents (.docx)
  • Text files (.txt)
  • And other common document formats

Notes​

  • Documents are processed and indexed to enable efficient retrieval during AI interactions.
  • The quality and relevance of AI responses depend on the quality and comprehensiveness of uploaded documents.
  • You can upload multiple documents to create a comprehensive knowledge base.
  • Consider organizing documents by topic or subject area for better AI performance.
  • Large documents may take longer to process and index.

Best Practices​

  1. Use Clear, Well-Structured Documents: Documents with clear headings, sections, and formatting will produce better AI responses.
  2. Keep Content Relevant: Only upload documents that are directly relevant to the questions users will ask.
  3. Update Regularly: Keep your knowledge base current by updating documents when information changes.
  4. Test Responses: After uploading documents, test the AI responses to ensure it's retrieving and using information correctly.
  5. Combine with System Prompts: Use Additional System Prompts to guide how the AI should use the knowledge (e.g., citation requirements, tone, response length).