Product Tagging is a powerful AI-driven feature that automatically extracts detailed product attributes from your studio images, transforming visual content into structured, revenue-driving data. Using advanced computer vision technology, it identifies key details like material, fit, style, and more, eliminating manual tagging errors and dramatically reducing the time needed to enrich your product catalog.
Just a heads-up: Product Tagging is an add-on feature. If you're interested, reach out to your Customer Success Manager for more info.
Current Capabilities
Product Tagging currently supports:
Fashion products.
For the input image, Ghost mannequin images work best.
80+ pre-defined attributes across categories like Consumer, Design, Physical, Style, and Usage; custom attributes and categories can be added.
Product tags can be exported product-by-product; bulk exports are coming soon.
Setting Up Product Tagging With Schema Builder
Enabling Product Tagging
Before starting with the schema the first thing to ensure is that the Product tagging is enabled for your Studio. This is something you can verify on the Studio Settings, under the “Dreem.ai” connector.
Creation and Editing of Attribute Schema
Now that you have your add-on enabled you'll need to set up an attribute schema in Gamma. The schema acts as a blueprint that defines which attributes to extract and their acceptable values for specific product categories. The product category will be used to match schema and products together.
In order to do this, navigate to the Schema Builder in Gamma, available under “Attribute Schemas” in your studio settings.
Here you’ll find a list of existing schemas and will also have the ability to create new ones.
To create a new schema, you can click on “Add Schema”, in the top right corner of your screen, and to edit an existing one simply click “Edit” on the schema of your choice.
If you decide to create a new one, you’ll have to give it a name, select a specific workspace and category.
Please note that each category can only belong to one schema per workspace, but a single schema can cover multiple categories.
This might explain why some categories are not available if they are already used in another schema under the same workspace.
Building your Schema Attributes
It’s now time to start building your schema. To do this click on the “Attributes” tab and “Schema Builder” on the bottom left.
Let’s dive into the builder, don’t be afraid, it’s much simpler than it looks!
First at the top you’ll see that you can switch between the “Builder” or the “Tester” view, but for now let’s concentrate on the builder.
On the left side you’ll find a list of your attributes, roughly 80 by default but you can create as many as you want, and they’ll be categorised in 5 main categories: Consumer, Design, Physical, Style and Usage.
You also have a search bar to look for a specific attribute, and can add new attributes (which will then become available for any other new or existing schema).
If you want to add an attribute, simply tick its box and it’ll automatically be added on the schema visible on the right side.
In the following example we have selected the attributes “Fit Type” and “Gender Marketing”, so you can see on the schema the attributes in purple, with a list of the potential return values in blue.
To edit an existing attribute, you can click on the logo near the attribute, then click “Edit”.
From here you can rename the attribute, change its category or description and also modify the list of potential return values if needed.
For the “Returns” section, if you choose “Single Value” then only one of the values listed will be returned by the AI based on your product.
If you select “List of Values” you give a bit more freedom for the AI to choose multiple values from the list.
In both cases the values returned will be chosen from the defined list in “Potential Return Values”.
If you would like to give even more freedom to the AI and let it choose the value, then you can leave the “Potential Return Value” list empty.
Testing Your Schema
The Schema Builder includes a built-in tester that lets you validate your configuration before adding it to a live workflow.
To access it click on the “Tester” tab at the top.
Then simply choose a product code within Creative Force to see how the AI will tag it using your current schema.
Please note that the product must be part of the same workspace as the schema you are building.
You just need to add your product code to the search box in the top left, the product should be recognized by the system, then click “Generate Attributes” to start the testing.
If you have multiple images for the same product you can also select which specific one should be used for the test.
The analysis typically takes 1-2 minutes, and you can refine your attribute definitions based on the results.
If an attribute can't be identified from the image, it will be left blank rather than guessing incorrectly.
Adding Product Tagging to Your Workflow
Workflow Integration Requirements
Product Tagging requires Workflow Engine V2 and can be added as a sub-function to any step in your workflow builder.
The most common placement is post-final selection, to ensure only approved images are processed, but it can be positioned wherever best fits your production process before final delivery.
If you have multiple steps of product tagging in the same workflow the last one will just override the previous one.
In your workflow V2 builder, navigate to the step where you want to add tagging. Here click on the “Add” cloud automation.
Then choose Dreem.Ai.
And finally you can select “Product Tagging”.
Note: If you want to add the product tagging to a “Cloud Automation” step, there's currently a UI quirk where you need to add an additional sub-function after product tagging to properly save the configuration.
Reviewing and Managing Tagged Data
Once tagging is complete, you can access the results from the product slide-in by clicking here on “View Attributes”.
From the new window you can check each product attribute, as well as editing them manually if needed.
If you want this is also where you can export them as JSON or CSV.
You’ll currently be able to do this only product by product but bulk download is on the way!
One last thing to note, from here you can also access the version history of the changes applied to the product attributes by clicking on “Change History” in the top right.
Here you can see the different versions that the attributes went through and decide which one you want to download.