Catalogs
Overview
Catalogs allow you to upload structured data about your products or services so that your AI agents can answer detailed questions about pricing, availability, features, and more. Unlike documents (which contain unstructured text), catalogs organize information into rows and columns, enabling both semantic search and structured filtering. You can have up to 3 catalogs, each with up to 3,000 items.
How to Access
- Navigate to Tools > Aurora AI > Knowledge Base > Catalogs tab
- Required role: Chatbot Admin
Interface Overview
The Catalogs tab displays:
- A header showing the current count of catalogs (e.g., "Catalogs (1 / 3)")
- An Import button (hidden when 3 catalogs already exist)
- Catalog cards showing name, description, and processing progress
- When you click on a catalog card, a detailed table view of its items appears with search functionality
If no catalogs exist, an informational panel explains what catalogs are and how to get started.
Features & Actions
Import a New Catalog
What it does: Creates a new catalog by importing data from a CSV file using the built-in CSV importer.
Steps:
- Click the Import button
- Enter a Name for the catalog
- Enter a Description for the catalog
- Click Select File to open the CSV importer
- Upload your CSV file and map the columns as needed
- Once the file is processed, you will see the number of items to be imported
- Click Import to upload the data
Important notes:
- Maximum of 3,000 rows per import
- Maximum of 3 catalogs per account
- The importer supports multiple languages (English, Spanish, Portuguese)
- Processing happens in the background with a progress indicator
- The CSV importer automatically maps column headers; you can adjust the mappings before importing
View Catalog Items
What it does: Opens a detailed table view showing all items in a catalog with their properties.
Steps:
- Click on a catalog card
- The table displays all items with columns based on the catalog's data structure
- Use the search box to filter items
- Click Search to apply the filter
- Click the back arrow to return to the catalog list
Edit a Catalog
What it does: Changes the name and description of an existing catalog.
Steps:
- Click the menu icon (three dots) on the catalog card
- Select Edit
- Modify the Name and/or Description
- Click Submit
Delete a Catalog
What it does: Permanently removes a catalog and all its items.
Steps:
- Click the menu icon (three dots) on the catalog card
- Select Delete
- Confirm the deletion in the dialog by clicking Delete
Important notes:
- This action cannot be undone
- If an AI agent was using this catalog, it will no longer have access to its data
Search Within a Catalog
What it does: Filters catalog items by keyword.
Steps:
- Open a catalog by clicking its card
- Type a keyword in the search box
- Click Search
- To clear the search, click the clear button
Fields Reference
| Field | Description | Required | Notes |
|---|---|---|---|
| Name | Display name for the catalog | Yes | Must not be empty |
| Description | Brief description of the catalog contents | Yes | Must not be empty |
| CSV File | The data file to import | Yes | Max 3,000 rows |
Important Notes
- Maximum of 3 catalogs per account
- Maximum of 3,000 items per catalog import
- Catalog processing happens in the background; progress updates are shown in real time via a progress bar
- When processing reaches 100%, the catalog status changes to "ready"
FAQ
Q: What types of data can I put in a catalog? A: Any structured data with columns and rows works well. Common examples include product catalogs with name/price/description, service menus, inventory lists, and course catalogs.
Q: How does the AI use catalog data differently from documents? A: Catalogs support structured filtering (e.g., "show me products under $50") in addition to semantic search. Documents only support text-based semantic search.
Q: Can I update items in a catalog after importing? A: Currently, you can view and search catalog items. To update the data, delete the existing catalog and re-import with the updated CSV file.
Q: What file format does the importer accept? A: The importer accepts CSV files. You can map columns during the import process to match your data structure.