Attributes & Enrichment

Understanding Attributes in Merchkit

Learn what attributes are, where to find them, and why they're central to both your data structure and AI enrichment setup.

On this page

Understanding Attributes in Merchkit

An attribute is both a column in your product table and the configuration that tells Merchkit's AI what to generate. This dual role is central to how Merchkit works: your data structure and your enrichment instructions are one and the same.

What Are Attributes?

Think of attributes as the fields that describe your products. In Merchkit, each attribute can be:

  • A product property (title, description, price)
  • A data field you've created (brand, material, fit_type, care_instructions)
  • An AI-generated field (SEO title, short description, category recommendation)

Every attribute has a name, a field type (Text, Number, Image, Formula, etc.), and optionally an AI instruction that tells Merchkit how to generate or enrich its value.

Where to Find Attributes

Open Workspace → Attributes in your sidebar.

[SCREENSHOT: Workspace menu with Attributes option highlighted]

You'll see the Attributes page, which displays a table of all your configured attributes. At the top, you can switch between entity types: Products, Categories, Vendors, Images, and Sources. Each entity type has its own set of attributes.

[SCREENSHOT: Attributes page with entity type dropdown showing Products selected]

Entity Types: Attributes Aren't Just for Products

Merchkit lets you define attributes for multiple entity types:

  • Products — your main product data
  • Categories — hierarchy and metadata for your product categories
  • Vendors — supplier information and details
  • Images — metadata about product images
  • Sources — information about where your product data comes from

You configure attributes for each entity type separately. This guide focuses on Products, but the same principles apply to the others.

How Attributes Control AI Enrichment

Each attribute has several configuration settings that work together to tell the AI what to generate:

Field Type — Determines what kind of data the attribute holds (text, number, image reference, etc.)

Use AI toggle — Enables or disables AI generation for this specific attribute

Product Data — Controls which product fields (title, description, price, etc.) the AI can see when generating this attribute's value

Image Source — Tells the AI which product images to analyze (if the attribute requires image analysis)

Prompt — Your custom instruction to the AI. This is how you specify exactly what to generate. Supports dynamic references to other attributes using {{ }} tags.

Acceptable Values — A comma-separated list that constrains the AI's output to only those values

Together, these settings define what the AI sees, what the AI does, and what the AI is allowed to return.

A Key Insight: The Attributes Page Is Your Enrichment Configuration

Unlike many tools where data structure and enrichment are separate, in Merchkit they're unified. When you configure an attribute, you're simultaneously:

  • Defining a column in your product table
  • Setting up an enrichment instruction (if Use AI is enabled)
  • Specifying the AI context, constraints, and format

This means: attributes must be configured before running enrichment. You can't enrich a field that doesn't exist yet.

Field Types (Brief Overview)

Merchkit offers several field types for your attributes:

  • Single Line Text — Short text values (product name, SKU, color)
  • Number — Numeric values (price, inventory count)
  • Image — Product images or image URLs
  • Formula — Computed values (based on other fields)
  • Reference Lookup — Links to related records (e.g., category)
  • Collection Reference — Multiple linked records
  • Record Reference — Single linked record

For detailed information about field types, see Field Types.

What You'll Do Next

In the next articles, you'll:

  1. Learn how Merchkit auto-creates attributes when you import from Shopify or another platform
  2. Walk through the Configure Attribute panel step-by-step
  3. Write effective AI prompts
  4. Set up acceptable values to constrain AI output

Next: Quick Start: Auto-Created Attributes from Integrations or Configuring an Attribute for AI Enrichment