Learn the core building blocks of Merchkit — Products, Sources, Attributes, Categories, Integrations, and Channels — and how they relate to each other.

Key Concepts & Terminology

Before you start working in Merchkit, it helps to understand the core concepts you'll encounter. This article explains what each one means and how they connect to each other.

For a complete glossary with synonym mappings, see Glossary & Terminology.

Products

A Product is the central record in Merchkit. Each product represents one item in your catalog — a sofa, a rug, a mirror. Products live in the Products table, which is the main view you'll see when you log in.

Each product has attributes (data fields like Title, Description, Price, Color), can have sources attached to it (documents, URLs, or text that give the AI context), and can have images.

Products can also have parent-child relationships. A parent product might be "Arya Machine Woven Rug" and its children are the individual size/color variants.

Sources

A Source is any piece of reference material you attach to a product to give the AI more context for enrichment. Sources can be:

URLs — a link to a product page, spec sheet, or supplier listing
Text — raw product descriptions, specification data, or notes pasted directly
Documents — uploaded files such as PDFs, Excel spreadsheets, Word docs, ODT, or RTF (up to 25MB per file)

Sources are attached at the product level. When you click on a product's "Data Sources" cell, a panel opens where you can add URLs, text, or drag-and-drop files. The Sources page in the sidebar shows a flat list of all sources across your entire catalog.

Sources are different from your initial CSV import. The CSV import creates your products and populates their basic attributes. Sources provide additional context that the AI uses during enrichment — think of them as the raw material the AI reads to generate better output.

Attributes

An Attribute is a data field on a product — like Title, Brand, Color, Price, SEO Description, or Carton Width. Attributes are the building blocks of your catalog.

In Merchkit, attributes are more than just static fields. Each attribute can be configured with:

A field type — what kind of data it holds (text, number, image, formula, reference lookup)
An AI toggle — whether the AI should generate values for this attribute
A prompt — instructions that tell the AI exactly what to generate and how
Acceptable values — a constrained list that forces the AI to return only specific values (e.g., for a Color attribute: "Red, Blue, Green, Black, White")
Product data context — which other data the AI should reference when generating (product attributes, data sources, images, or web search)
Required values — which other attributes must have data before this attribute can be generated

Attributes are managed from Workspace → Attributes in the sidebar. They can be scoped to different entity types: Products, Categories, Vendors, Images, or Sources.

The Attributes page is the control center for enrichment quality. Every "why did the AI generate this?" question traces back to how the attribute was configured here.

Integrations

An Integration is a connection to an external platform for importing and exporting data. Integrations are data pipes — they move product information between Merchkit and tools like Shopify, BigCommerce, Salsify, Akeneo, or WooCommerce.

Each integration has:

Import mapping — how fields from the external platform map to Merchkit attributes (e.g., Shopify's "Title" field → Merchkit's "Title" attribute)
Export mapping — how Merchkit attributes map back to the platform's fields
Import filters — rules that let you import specific subsets of products (e.g., only products in a certain collection or category)

A useful shortcut: when you import from a platform like Shopify, Merchkit automatically creates attributes from the platform's existing fields and metafields, giving you a head start on attribute setup.

Integrations are managed from Workspace → Integrations in the sidebar.

Channels

A Channel is a sales destination where your products are listed — like Wayfair, Amazon, Walmart, Google Merchant, Nordstrom Rack, or even AI shopping agents like ChatGPT.

Channels are fundamentally different from Integrations. While Integrations move data in and out, Channels generate content. Each channel has:

A dedicated grid view with its own set of channel-specific attributes
Pre-loaded attribute templates with default AI prompts tailored to that channel's requirements
Category-specific views for channels that have different attribute requirements per product category (e.g., Wayfair's "305 – Sofas" view)

When you open a channel, you see a purpose-built view showing only the attributes that channel cares about, with AI prompts already configured to generate values in the format that channel expects. You then select your products, click Generate, and Merchkit creates channel-ready content you can export for submission.

Channels are listed in the Channels section of the sidebar.

How they all connect

Here's the flow, from raw data to channel-ready content:

You import your catalog (via CSV or an Integration) → this creates your Products
You add Sources to products for additional AI context
You configure Attributes with AI prompts, field types, and acceptable values
You run Generate on your products → Merchkit's AI enriches your catalog attributes
You review and refine the output
You open a Channel view → Merchkit generates channel-specific content from your enriched catalog
You export the channel data for submission

Categories organize your products throughout this process, and Integrations let you sync data with your existing platforms at any stage.

Next step: Read From Zero to Optimized Catalog in 7 Days to see this flow as a step-by-step checklist for your first week.