Skip to main content

​
How It Works

Knowledge Base uses Retrieval-Augmented Generation (RAG) to give your agent access to your documents during conversations. When a caller asks a question, the agent searches your uploaded documents for relevant information and includes it in its response.

​
Setting Up Knowledge Base

1

Navigate to Agent

Go to Agents → select your agent → Knowledge Base tab.
2

Upload Documents

Supported formats:
  • PDF (.pdf)
  • Word Documents (.docx)
  • Text Files (.txt)
Drag and drop or click to upload. Multiple files can be uploaded at once.
3

Enable RAG

Toggle RAG Enabled to activate knowledge base retrieval during calls.
4

Test

Use the Query feature to test retrieval:
  • Enter a sample question
  • View the retrieved chunks and sources
  • Verify the agent finds the right information

​
How Retrieval Works

  1. Uploaded documents are split into chunks and converted to vector embeddings using OpenAI
  2. Embeddings are stored in a local file-based index (LlamaIndex)
  3. During a call, when the agent decides it needs information, it:
    • Converts the user’s question to an embedding
    • Finds the most similar chunks from the index
    • Injects the relevant text into the LLM prompt as context
  4. The LLM uses this context to generate an informed response

​
Best Practices

  • Keep documents focused — upload specific product docs, not entire websites
  • Update regularly — re-upload when your information changes
  • Test queries — verify retrieval accuracy before going live
  • Combine with system prompt — tell the agent when to use the knowledge base vs. general knowledge

​
Limitations

  • Documents are stored on the local filesystem (not S3)
  • Index rebuilds when files are added/removed (may take a few seconds for large documents)
  • Currently uses OpenAI embeddings — requires OPENAI_API_KEY
  • No support for spreadsheets, images, or audio files