If your knowledge base is larger than the 37,000 character limit in the project settings, you can use RAG (Retrieval-Augmented Generation) to connect your AI Agent to a bigger external knowledge source.
This allows the agent to fetch relevant information dynamically instead of relying only on the built-in knowledge base.
Note:
You can create up to 5 knowledge bases, and each one can include up to 5 files.
Important: Preparing Your Knowledge Base File
Your knowledge base file should be in Markdown (.md or .mdx) format.
Markdown is a simple way to structure content using titles and headings.
Use clear titles and headings (e.g., # Title, ## Subtitle) — the system uses them to organize the content.
If you have tables, split them into smaller sections and add a title for each.
Any content without titles will be treated as one block, which may affect accuracy and increase token usage.
Example:
# Title 1
## Subtitle 1
- option 1
- option 2
# Title 2
## Subtitle 2
- option 1
- option 2
Step-by-Step: How to Set Up RAG for Your AI Agent
1. Log in to your Brightcall account.
2. Go to Account Settings.
3. Open the RAG Settings tab.
4. Click on Create Knowledge Base.
5. Add a Name (Description is optional), then click Create.
6. From the Actions column, click on the three dots (⋯) and select Manage.
7. Go to Documents, then click Upload File and upload your Markdown (.md/.mdx) file.
8. Go back to your AI Agent Project where you want to add the knowledge base From Project Settings, open the Calls tab.
9. Scroll down and click on Show Advanced Settings.
10. Scroll down to AI RAG Knowledge Bases, and select the knowledge base you created from the dropdown.
11. Click Save.
Need help? Contact us at [email protected]