Skip to main content
SEO and acquisition apps

Shopify Caps Storefront Filters at 25 Per Store — The Full List of Search & Discovery Limits

Shopify's free Search & Discovery app has published ceilings: 25 filters per store, no filters on collections over 5,000 products, 100 filter values shown. Some search features also depend on locale. Here is every documented limit and what the paid apps actually lift. Researched August 31, 2026.

AIEC App Lab AI Editorial TeamPublished Updated Read 11 min#SEO and traffic#App comparisons#Shopify blog operations#For beginners

What you will learn

On this pageContentsCloseOpen

Every growing catalog hits the same wall eventually: you want one more filter, and the app will not let you add it.

Shopify's free Search & Discovery app caps filters at 25 for the entire store. Not 25 per collection. Twenty-five, total. A Shopify Community thread opened in January 2024 collected eleven replies from merchants explaining that 9,000 products need more than 25, and that competitors run 50, 60, even 300 filters.

There are other ceilings. Collections holding more than 5,000 products display no filters at all. A single filter shows a maximum of 100 values to shoppers. And several search features are locale-dependent in ways the marketing pages do not mention — semantic search requires the Grow, Advanced, or Plus plan and is not available in every locale.

This article lists every numeric limit Shopify publishes, then covers what the paid apps actually lift, restricted to what their own documentation confirms. Researched August 31, 2026.

The short answer

If you are under about 1,000 products, no collection exceeds 5,000 products, your filter axes fit inside 25, and you sell in one currency and one language, the free Search & Discovery app is enough. It costs nothing and it is a first-party integration.

When merchants do hit a wall, it is almost always the 25-filter cap or the 5,000-product collection. Which one you hit determines which app you should look at.

Paid apps meter on three different things. Product count (Smart Product Filter & Search, Searchanise), monthly GMV (Boost AI Search & Filter), and monthly sessions (Fast Simon). A high-revenue store with few SKUs and a low-revenue store with many SKUs will find opposite apps cheapest.

"Unlimited" usually is not. Boost's own documentation caps filter options at 30 per filter tree. That beats 25, but it is not uncapped.

Every documented limit

These are the numbers written in the Help Center. They are not negotiable.

SubjectLimit
Total filters25 per store (standard plus custom combined)
Filter source reuseEach source can be used once per store
Large collectionsCollections over 5,000 products display no filters
Large result setsSearches returning over 100,000 results display no filters
Filter values shownMaximum 100 per filter on the storefront
Filter groupsMax 200 unique values per group, 1,000 groups per store
Values shown in the appMaximum 1,000
Unique tag valuesMaximum 5,000
Unique product option and attribute valuesMaximum 1,000
Synonyms5 words per synonym, 20 synonyms per group, 1,000 per store
Product boosts10 search terms per boost
Recommendations10 complementary products, 10 related products
Analytics retentionLast 30 days inside the app
Index propagationCan take 24 to 48 hours

The six standard filters are Availability, Category, Price, Product type, Tags, and Vendor. Custom filters are built from product options, metafields, or metaobjects. The six standard filters consume the same 25 slots. Turn on Price and Tags and you have 23 left.

The reuse rule catches people out too. Once Price is active as a filter source, it is greyed out — you cannot build a second, differently-framed price filter.

The limits that are not numbers

The non-numeric constraints tend to cause more operational pain.

Multi-currency stores lose the price filter. The Help Center states that with international pricing or multiple currencies enabled, the price filter does not display when a shopper selects any currency other than the store's default. It describes this as a current limitation of the price filter.

Tag and vendor filter values cannot be translated. As a result, the product tag filter is only shown to customers shopping in the store's default language, and vendor filter values are always based on the default language. If your storefront filtering is built on tags, that design does not survive going multilingual. Market-specific language variants (Canadian French versus France French, say) also fall back to the general language for filter values.

Boolean logic is mostly fixed. Values across different filters combine with AND; values within the same filter default to OR. Only product tags, metafield list types, and metaobject reference list types can be switched to AND.

You cannot reorder search results. There is no setting for it. The Help Center notes that Shopify runs ongoing tests on search ranking, and that during a test some customers may see slightly different result ordering than others.

Two setup requirements are easy to miss: using Search & Discovery requires Shopify Network Intelligence to be enabled in your privacy settings, and staff need the Products, Online Store Search and Navigation, and Reports permissions.

Search behavior is locale-dependent

This is the part most comparison articles skip, and it changes what you should expect from the free app.

Semantic search — AI understanding of meaning — is gated three ways. It requires fewer than 200,000 products, requires the Grow, Advanced, or Plus plan, and does not apply to predictive search at all. On top of that, it is not supported in every locale; the Japanese locale, for example, is explicitly excluded.

Typo tolerance is also locale-dependent. Where it is supported, it corrects terms differing by one letter or with two letters transposed — but the first four letters must be entered correctly, and it does not apply to tags at all. So chocollate will not match a product tag no matter what.

Underneath all of it, Shopify's storefront search matches complete words. The Help Center states plainly that a partial word such as artich might not return products with artichoke in the title. For languages that do not separate words with spaces, Shopify substitutes trigram matching, which is enabled only for those locales.

The searchable fields are worth memorizing. For products, search covers body, product type, tag, title, variant barcode, variant SKU, variant title, and vendor. But predictive search — the dropdown as customers type — only covers title, product type, variant title, and vendor. Body and tags are excluded. Most "I searched for a word that's in the description and got nothing" tickets trace back to exactly this.

SKU search has a quirk of its own: exact matches work, but partial matching only functions when the code contains hyphens. abc1-888 is findable by searching abc1; abc1888 is not.

And using search syntax (OR, -, "...", title:) disables typo tolerance, predictive search, semantic understanding, product boosts, and synonyms. Your most sophisticated shoppers get the least tuned experience.

Deciding whether the free app is enough

The free app is enough for stores up to roughly 1,000 products where no collection exceeds 5,000 items, filter axes stay within 25, and you operate in a single currency and language.

You will hit the ceiling if you sell parts or tools where axes multiply (manufacturer, fitment, size, material, standard), if you keep products in one large collection, if you need the price filter visible in multiple currencies, or if you want tag-based filtering across languages.

The Community has receipts. A June 2023 post documents filters vanishing the instant the 5,000th product was added to a collection on the Dawn theme. A December 2025 thread found that a site displaying filters across 25,000 products was running an external search service.

The paid apps

App Store listings as of August 31, 2026.

Search & DiscoverySmart Product Filter & SearchSearchanise Search & FilterBoost AI Search & FilterFast Simon AI Search & Filters
DeveloperShopifyGlobo.ioSearchaniseBoost CommerceFast Simon
Billing unitFreeActive productsActive productsMonthly GMVMonthly sessions
Free planAll features freeDev stores onlyUp to 25 productsNone100 sessions/month
Cheapest paidNone14 USD/mo19 USD/mo29 USD/mo39.99 USD/mo
Free trialN/A14 days14 days21 daysNot stated
Rating2.74.94.84.84.8
Reviews4632,2331,0701,500286
Built for ShopifyNoYesNoNoNo

Three observations.

The first-party app's 2.7 rating is the feature gap showing up as a score. It carries 239 five-star reviews against 90 one-star reviews (19%). A July 2, 2026 review from a US merchant criticizes the 25-filter cap in blunt terms. That is an unusually low rating for a free first-party app, and it maps directly onto the constraints above.

Because the billing units differ, the cheapest app changes with your shape. 5,000 products and modest revenue favors Searchanise (7,500 products at 39 USD/month). 200 products and high revenue is penalized by Boost's GMV-linked pricing. On Fast Simon's session metering, a November 6, 2025 review reports that Google bot traffic counted toward the session total and pushed the store into a higher tier.

Only Smart Product Filter & Search carries the Built for Shopify badge among the four paid apps. The first-party Search & Discovery does not carry it either. The badge reflects performance, design, and integration standards — not feature breadth.

What each app verifiably lifts

Restricted to what vendor documentation confirms. "Unlimited" in marketing copy frequently is not.

On the 25-filter cap, Smart Product Filter & Search and Searchanise both advertise unlimited filter trees on every plan. Searchanise's FAQ states explicitly that custom filter trees, range sliders, merchandising campaigns, analytics, and sync frequency are unlimited regardless of plan.

Boost AI Search & Filter is not unlimited. Its support documentation specifies 30 filter options per filter tree and 1,500 filter values per option. That clears 25 and 100, but it is a ceiling. The number of trees is also plan-dependent: 5 on Launch, 20 on Convert, unlimited on Accelerate. Exceed your plan's product count and each tree is reduced to displaying a single option, with sync degrading and eventually stopping.

On the 100-value display cap, Boost's 1,500 is the only figure confirmed in official documentation. The other three do not publish a value ceiling.

On the 5,000-product collection limit, Boost's company blog names it explicitly as a problem they solve. However, the same post describes Shopify as allowing "4 standard plus 20 custom" filters, which does not match the current documentation (6 standard types, 25 total). Stale content is clearly present, so that post should not be treated as a primary source. For the other three, plan tiers extending from tens of thousands to 2,000,000 products imply support, but it is not confirmed.

Separately, the Help Center notes that when a third-party search app powers your storefront search, standard search behavior and product visibility settings may not apply — including Unlisted product status. Products you believe are hidden from search can surface in app-powered results. Verify this after installing.

Choosing by situation

Under 1,000 products with fewer than 25 filter axes

Stay on Search & Discovery. Spend the effort on synonyms and product boosts instead, and work through the "searches with no results" report in analytics. That returns more than a subscription would.

You have hit only the 25-filter cap

Smart Product Filter & Search is the cheapest entry at 14 USD/month and is the one paid app in this set carrying the Built for Shopify badge. It also handles fitment-style search for parts catalogs.

Tens of thousands of products, or a volatile catalog

Searchanise offers the widest product range, up to 2,000,000. Its FAQ commits to filter features being unlimited regardless of plan, so downgrading does not silently remove capability.

Large revenue, filters tuned per collection

Boost AI Search & Filter fits. Note that pricing tracks GMV, so cost rises as you grow. Know about the 30-options-per-tree cap before you commit.

Headless, image search, or POS alongside

Fast Simon provides Hydrogen and React SDKs and lists POS integration. Session-based billing makes cost harder to forecast, particularly given the bot-traffic report above.

Pre-install checklist

  • How many filters are you using now? The six standard types count toward the 25
  • Does any collection exceed 5,000 products?
  • Does any filter have more than 100 values?
  • Do you have multiple currencies enabled? The price filter disappears for non-default currencies
  • Are you filtering on tags while selling in more than one language?
  • Have you separated the search problem from the filtering problem?
  • Have you tested actual customer queries, including in every locale you sell in?
  • Which of product count, GMV, or sessions grows fastest for you? Match the billing unit to it
  • Have you confirmed Unlisted products stay out of app-powered search results?
  • Is Shopify Network Intelligence enabled in your privacy settings?

In summary

Shopify's storefront filtering comes down to three numbers: 25 filters per store, 100 values per filter, 5,000 products per collection. If you are inside all three, the free app is sufficient.

  1. Count your current filters and the size of your largest collection
  2. Identify which ceiling you have actually hit — filter count, collection size, or value count
  3. Decide whether product count, revenue, or sessions grows fastest for you, then pick an app whose billing unit matches
  4. Use the free trial to test real queries and your real filter configuration, in every locale you sell in

Limits and pricing change. Verify against the Shopify Help Center and each app's App Store listing before installing.

RelatedShopify's Abandoned Checkout Email Never Reaches Anyone Who Only Filled a Cart

Shopify only counts a checkout as abandoned once a shopper enters contact info and stalls for ten minutes. That excludes every cart-only exit, every POS and Shop app checkout, and anything older than three months. Plus the conditions that silently block delivery, the opt-in you cannot undo, and how three recovery apps bill on three different units. Researched August 31, 2026.