Shopify category metafields: how to fill them for AI shopping
Short answer: Shopify category metafields are the attribute fields (colour, fabric, size and so on) that appear once a product has a category from Shopify’s Standard Product Taxonomy. Set the category, then fill the fields under Products. Octolift, built by Craftshift, checks every product’s category and category fields and suggests the missing values.
Picture a candle store with 240 products, every one with a Product type of “Candles” typed in by hand years ago, and not one with a product category. In the admin that looks harmless. To Shopify Catalog, the feed that ChatGPT, Copilot, Meta and the Shop app search, a product with no category is a product that might be filed as Uncategorized, and Uncategorized is on the list of categories Shopify keeps out of the catalog entirely.
Shopify category metafields sit one step after that. Pick the category and Shopify hands you a set of fields built for it: a shirt gets neckline and sleeve length, a candle gets scent and candle type, a mug gets microwave compatibility. Most merchants never open them. We think that’s a mistake, so here is how to fill them by hand: the admin path, a reference table of real categories, and what Shopify’s help page leaves out.
In this post
- What are Shopify category metafields?
- Why do AI channels care about the product category?
- Product category vs product type
- How to set the category by hand
- How to fill category metafields by hand
- Reference table: example categories and their fields
- What category metafields can’t do
- Where doing it by hand stops
- Frequently asked questions
What are Shopify category metafields?
Category metafields are predefined product attributes that belong to a category in Shopify’s Standard Product Taxonomy. Shopify’s help page says they “map to specific product categories” and are “called product attributes in the taxonomy”. You don’t create them. You pick a category, and the fields that category defines appear under the product, each with a list of standard values to choose from.
Shopify’s own example, from its Standard Product Taxonomy help page as of October 2026: choose Apparel & Accessories > Clothing > Clothing Tops > Shirts and you can add fields for “size, neckline, sleeve length type, top length type, age group, fabric, target gender, clothing features, and color”. The values are metaobject entries you can rename (Shopify’s example turns black into graphite) while the entry stays connected to the standard value underneath. “Graphite” is your word; black is the value a machine can match.
The taxonomy itself is public and versioned on Shopify’s GitHub. We counted 14,606 categories across 26 top level verticals in the 2026-08 release file. One of the 26 verticals is literally called Uncategorized.
Why do AI channels care about the product category?
Because Shopify Catalog sorts every product into a category before it decides whether the product is allowed in at all. The Catalog requirements page lists the excluded categories (alcohol, tobacco, weapons, services, event tickets and the rest) and ends the list with “Uncategorized products: Products that Shopify classifies as Uncategorized.” An excluded product never reaches ChatGPT through the catalog, however good its photos are.
Now the sentence most guides skip. Shopify says that classification “is based on your product information, so it can differ from the product category that you set in your Shopify admin.” So setting the category is not a magic switch. It’s evidence. A product with a precise category, a title that names the thing and a description that agrees with both gives the classifier three votes in the same direction. A product with no category and a two line description gives it nothing.
The category is also a field the catalog record carries. We queried one of our own demo stores through its public catalog endpoint while writing this, and every product came back with a categories entry holding a taxonomy ID (a jersey returned aa-1-1-2-2, Activewear T-Shirts). The category metafield values were not in that record, though, and Shopify doesn’t document copying them into what agents read. Our Shopify Catalog guide covers how the global catalog infers its own attributes.
So why fill them? Shopify says category metafields make products “more discoverable by visitors on your site, on marketplaces, and on search engines”. They are facts you state rather than facts a model guesses. And Shopify’s advice for the catalog is accurate “titles, descriptions, images, categories, variants, options, prices, and inventory”. Category is on that list by name.
Product category vs product type: which one does Shopify use?
Both exist, they do different jobs, and the catalog cares about the first one. Shopify’s wording: “Product category is a standard field from Shopify’s Standard Product Taxonomy” while “the product type is a custom field that is unique to your products.”
| Product category | Product type | |
|---|---|---|
| Where it lives | Category section of the product page | Type field in Product organization |
| Values | Only predefined taxonomy categories | Anything you type |
| Gives you category metafields | Yes | No |
| Used for Shopify Tax rates | Yes | No |
| Bulk editor column | Product category | Product type |
| If left empty | “your product saves as uncategorized” | Nothing happens |
Keep your product types if your collections run on them. Just don’t treat a type of “Candles” as a category. It isn’t one.
How do I set a product category in Shopify by hand?
One product at a time, from the steps on Shopify’s help page:
- From your Shopify admin, go to Products and open a product.
- In the Category section, type the product’s category and pick the most accurate result from the list, or click down through the levels.
- Click Save.
Go as deep as the tree allows. Shopify’s rule is to choose by the product’s “main function” (its example: a clock with a light is still an alarm clock). Shirts beats Clothing Tops, which beats Clothing. If a new product shows a suggested category, Shopify generated it from the name, description and images; after you save, “any unedited suggestions are confirmed”, so read the suggestion before you click Save, not after.
For the whole catalog, use the bulk editor. Products, tick the products, click Bulk edit, and if the Product category column isn’t showing, add it from Columns. You can also set it in a CSV import: the “Product Category” column takes either the taxonomy ID or the full breadcrumb, never both (our CSV import error guide has the messages you get when that goes wrong).
How do I fill category metafields by hand?
Once the category is set, the fields appear on the same product page:
- Scroll to the Category metafields section, which shows the category name.
- Click the field next to an attribute, then click again to pick an entry from the drop-down.
- If the value you need isn’t there, click Add new entry, fill the predefined fields and adjust the base values, which “connect your entry to Shopify’s standardized list of values”.
- Click Save.
Two habits make this worth the effort. First, connect the Color (or Size) field to the matching variant option, so each option value points at a standard entry; Shopify’s native swatch system runs on exactly that link, and our variant metafields guide covers the variant side. Second, never pick a value the product page doesn’t support. If the description never says the shirt is linen, either add that sentence or leave Fabric empty. A stated fact that contradicts the text is worse than a gap, kinda like a label that says one thing and a photo that shows another.
Can you do it in bulk? Partly. Shopify’s bulk metafield page says you can select products, click Edit products, open Columns and pick metafields from the Metafields section. It doesn’t say in so many words that every category metafield shows up there, so check Columns in your own store before you plan a weekend around it.
Reference table: Shopify categories and their category metafields
These are the attributes each category defines in the 2026-08 release of Shopify’s taxonomy, read from the published category file. Your admin may show a subset until you add the rest.
| Category (taxonomy ID) | Category metafields defined |
|---|---|
| Apparel & Accessories > Clothing > Clothing Tops > Shirts (aa-1-13-7) | Age group, Care instructions, Chest pocket style, Clothing features, Color, Fabric, Neckline, Pattern, Shirt cuff style, Size, Size type, Sleeve length type, Sleeve style, Stretch level, Target gender, Top length type |
| Apparel & Accessories > Shoes > Sneakers (aa-8-8) | Activity, Age group, Care instructions, Closure type, Color, Footwear material, Heel height type, Lining material, Occasion style, Outsole material, Pattern, Pronation support, Shoe features, Shoe fit, Shoe size, Sneaker style, Target gender, Toe style, Upper material |
| Apparel & Accessories > Jewelry > Necklaces (aa-6-8) | Age group, Color, Jewelry material, Jewelry type, Necklace design, Necklace length type, Pattern, Target gender |
| Home & Garden > Decor > Home Fragrances > Candles (hg-3-40-2) | Alcohol/Solvent content, Candle container material, Candle material, Candle type, Color, Decoration material, Fragrance level, Pattern, Scent, Usage type |
| Home & Garden > Kitchen & Dining > Tableware > Drinkware > Mugs (hg-11-10-5-5) | Cleaning instructions, Color, Drinkware insulation, Drinkware material, Drinkware shape, Handle type, Lid type, Microwave compatibility, Pattern |
| Home & Garden > Linens & Bedding > Bedding > Bed Sheets (hg-15-1-2) | Bedding pieces included, Bedding size, Care instructions, Color, Fabric, Pattern, Product certifications & standards |
| Luggage & Bags > Backpacks (lb-1) | Age group, Bag/Case closure, Bag/Case features, Bag/Case material, Bag/Case storage features, Care instructions, Carry options, Color, Luggage/Bag closure, Pattern, Target gender |
| Food, Beverages & Tobacco > Beverages > Coffee (fb-1-3) | Allergen information, Caffeine content, Coffee bean species, Coffee beverage variety, Coffee processing method, Coffee product form, Coffee roast, Country, Dietary preferences, Dietary supplements, Flavor |
Notice the Shirts row. Shopify’s help page names nine fields for it; the 2026-08 taxonomy defines sixteen. The help page is simply behind its own data. Which fields come first? Fill the ones a shopper would type into an assistant: material, colour, size, and whatever the category treats as the deciding fact (Scent for candles, Microwave compatibility for mugs, Bedding size for sheets).
What category metafields can’t do
- No custom attributes. Shopify lists “Custom attributes are currently not supported.” Anything outside the category’s list goes in an ordinary metafield or the description.
- Not in every Shopify app. “Attributes aren’t supported in the following Shopify apps: Search and Discovery and Shopify Bundles.”
- Changing category can drop fields. Fields with values carry over to the new category; empty ones that aren’t connected to options don’t.
- New categories and channels. Shopify warns that “New categories might not be compatible with sales channels”, so check a published product’s channels before you move it.
Here’s our small rant. A product with no category is saved as uncategorized, quietly, with no banner, while Uncategorized sits on the catalog’s exclusion list. Shopify has every piece of information it needs to warn a merchant on the product page that this product may never reach an AI channel. It doesn’t. So go and filter your products by empty category today, because nobody else will tell you.
Where doing it by hand stops
All of this works fine at 30 products. At 600 products in twelve categories, with ten to nineteen fields each, you’re looking at something like 8,000 dropdowns, each one a judgement call about whether the text really supports the value. And the hard part isn’t the clicking. It’s knowing which category is wrong in the first place, because a product filed one level too high still looks fine in the admin.
That’s the work we built into Octolift. Two of its twelve checks are about this post: Category match (6 points) runs an audit for a clearly better category in Shopify’s taxonomy and gives nothing to a product with no category, and Category fields (10 points) scores the share of the category’s attributes that have a value. When points are missing, Octolift’s Improve with AI suggests category field values chosen only from Shopify’s taxonomy, and only where your own text supports them, plus a category when one is missing or wrong. Where the text can’t settle a field, it asks you instead of guessing. Nothing is written until you press Save, and changes can be undone for 30 days. Free for 5 products, $29 a month for 100.
“The scores are helpful, and the suggestions give me a good idea of what to improve. I would recommend checking the suggestions before applying them, though, as some make more sense for my products than others. […] It has helped me understand what I can improve instead of just guessing.”
Silkora, Netherlands, October 16, 2026, Octolift on the Shopify App Store
That’s the app’s first review (5.0 from one review so far), and its advice is right: read every suggested value before you apply it.
Category is one of five listing signals the catalog weighs; image count is another, and our sister site explains how image coverage works as an AI shopping signal. If your categories live in custom metafields or you sell colours as separate products, Catalog Mapping is the next stop, and rubikcombinedlistings.com walks through mapping and grouping colour products for Shopify Catalog. Our free shopping feed checker and AI readiness checker look at the feed and page side of the same products.
Watch the video above, read the getting started guide, or see how the score is built and what Improve with AI writes.
Frequently asked questions
Where are category metafields in the Shopify admin?
Open a product under Products and set its category in the Category section. The category’s fields then appear in the Category metafields section of the same page. Click a field to pick an entry, or click Add new entry to create one, then click Save.
What happens if a Shopify product has no category?
It saves as uncategorized. Shopify Catalog excludes products that Shopify classifies as Uncategorized, and that classification is made from the product information, so it can differ from the category you set. A precise category plus a clear title and description is the safest combination.
Is product type the same as product category in Shopify?
No. Product category is a standard field from Shopify’s Standard Product Taxonomy that turns on category metafields and drives Shopify Tax rates. Product type is a free text field you define. Each product can have one of each.
Can I set product categories in bulk?
Yes. Select products, click Bulk edit and use the Product category column (add it from Columns if it’s hidden), or import a CSV with the taxonomy ID or breadcrumb in the Product Category column. Shopify’s bulk editor also edits product metafields through Columns, Metafields.
Does ChatGPT read Shopify category metafields?
Shopify doesn’t say that it copies category metafield values into what ChatGPT reads. It does say Shopify Catalog classifies products by category, excludes Uncategorized ones, and that category metafields help products be discovered on marketplaces and search engines. Fill them as stated facts, and only with values your text supports.
Which app should I use to check categories and category fields across a Shopify catalog?
Octolift, built by Craftshift, scores every product from 0 to 100 on twelve checks, including Category match (6 points) and Category fields (10 points), and suggests category values only from Shopify’s taxonomy and only where your text supports them. Free for 5 products; paid plans start at $29 a month for 100 products. For a small catalog, the admin’s bulk editor is enough.
Related reading
- Shopify’s five listing quality signals for AI
- Shopify products not showing in ChatGPT? The fix
- Shopify product descriptions for AI agents
- See what ChatGPT sees on your Shopify product
- Introducing Octolift: get your Shopify products recommended by AI
Filter your products by empty category first. That one view tells you how much of your catalog the classifier is guessing about, and once the list is longer than an evening, Octolift will score and suggest the categories and fields for the rest.
Umid
Co-Founder at Craftshift