Extract tables with Collections (Beta)

Collections are a property type that can be used to extract tables in your data, and pull their values into separate fields. Extracted properties can then be exported as structured data, or used as inputs on a project level.

To create a collection:

1. Add a new property with the Collection type

2. Select an AI Tool and an Input for the property
3. Add the properties (columns) that you'd like to extract. Go will be able to infer most properties within a file based on their name

3. Click on the collection items to review the extracted items

These values can then be referenced at the top level of the project. For example, after extracting invoice line items from a dataset of invoices, I can use "Invoice/Price" and "Invoice/Sales Tax" as inputs in a "Total sales tax" property:

🚧

Collections are read-only

The data extracted using collections is currently read-only. Any changes to the fields in a collection will not be persisted.