Configure enrichment

Choose Lily's sources, what she collects, the allowed or excluded sites, the image filters and the settings templates.

The enrichment settings decide where Lily looks and what she keeps. They are found in the enrichment step of the workflow Marcus proposes.

Open the settings

From the web: searches the web for data, images and documents.
From images: reads the images already attached to the product. No settings.
From linked documents: reads the documents already linked to the product. For a PDF catalog, only the pages where the product was spotted.

Add a source adds a method to the step. Each method can be turned on or off.

Product data: attributes and specifications.
Images and Documents: tick them according to your needs.
Extract attributes from collected images and Extract attributes from collected documents.
Maximum documents read per product: 4 by default, 20 at most.

These choices apply to all the step’s methods.

Accepted types: Packshots, Lifestyle images, Text on images.
Exclude images with watermarks or logos and Exclude amateur photos (second-hand products).
Minimum size in pixels and Max. file size (MB), 10 MB by default.
Image sources: All sources, Only these sources or Exclude these sources, with the list of addresses.

In Aggregation rules, describe how to decide between conflicting values, for example by favoring the brand’s website.

Your settings are saved automatically in the workflow.
Settings template: Save creates a reusable preset, Load applies it.
On the proposal card, Keep these settings for future runs remembers them for your next requests.

More documents read means a higher cost: each document read costs 2 credits.
A shared template can be used by other workflows: updating it changes them too.
The From a file and From a connected source methods are not available yet.