Access the Knowledge Base by selecting an agent from the Agents list and clicking the Knowledge Base tab.
How it works
Compass uses a technique called Retrieval-Augmented Generation (RAG) to connect your documents to the AI model:- Chunking — Your documents are split into smaller, manageable sections
- Embedding — Each section is converted into a numerical representation using AI
- Storage — These representations are stored in a vector database for fast retrieval
- Retrieval — When a user asks a question, the most relevant sections are retrieved and provided as context to the AI model
- Response — The AI generates a response grounded in your approved content
Source types
You can add two types of sources to your Knowledge Base:PDF Documents
Upload PDF files such as prescribing information, clinical study reports, product information leaflets, and approved marketing materials.
Website URLs
Add web page URLs to scrape and index content from your existing websites, product pages, or online resources.
Uploading PDF documents
Select your file
Click the upload area or drag and drop a PDF file. The file must be in PDF format and no larger than 10 MB.
Add metadata
Fill in the document details:
- Name — A descriptive name for the document
- Author — The document author or organisation
- Publication date — When the document was published
- Therapeutic area — The relevant therapy area
- Source type — The category of the document
Adding website URLs
Enter the URL
Paste the full website URL you want to index. The page content will be scraped and processed.
Add metadata
Fill in the source details — name, author, publication date, therapeutic area, and source type — just as you would for a PDF.
Document processing pipeline
After you upload a document or add a URL, it moves through a processing pipeline:| Status | Description |
|---|---|
| Pending | The document has been submitted and is queued for processing. |
| Processing | The document is being chunked, embedded, and indexed. |
| Indexed | Processing is complete. The document is now available to your agent. |
| Error | Something went wrong during processing. Check the document and try again. |
The Knowledge Base page automatically checks for status updates every 5 seconds while documents are processing. You’ll receive a toast notification when indexing is complete.
Managing documents
Once a document is indexed, you can manage it from the Knowledge Base list:- Edit metadata — Update the document name, author, publication date, therapeutic area, or source type
- Reindex — Trigger reprocessing if the source content has changed or if you want to refresh the indexed data
- Delete — Permanently remove the document from the Knowledge Base
Best practices
Choose authoritative sources
Choose authoritative sources
Keep documents up to date
Keep documents up to date
When product information changes (new indications, updated safety data, revised labelling), update or replace the relevant documents in your Knowledge Base promptly.
Use descriptive metadata
Use descriptive metadata
Accurate metadata helps you manage a growing knowledge base. Use clear, consistent names and always specify the therapeutic area and publication date.
Monitor for errors
Monitor for errors
Check back after uploading to confirm documents reach Indexed status. If a document shows an Error status, review the file and re-upload if needed.

