Guide · Structured data

Should a SaaS add SoftwareApplication structured data?

Updated 2026-11-07 · kagentx

Short answer

For a web app, WebApplication plus offers is enough. Google's documentation (checked 2026-11-07) requires name, offers.price, and either aggregateRating or review; it recommends applicationCategory and operatingSystem; ratings must come from real users and be visible on the page, and you cannot write your own. Our three sites: the MochiStamp and Gamma AI Slide homepages have a WebApplication with a price range and no rating; the kagentx homepage had only a WebSite and gained a WebApplication and the two prices on 2026-11-07. There is no evidence that these markups make AI recommend you more.

For a SaaS or web app, SoftwareApplication or WebApplication?

For a product used in the browser, write WebApplication. In schema.org WebApplication is a subtype of SoftwareApplication and inherits all its properties; Google's Software app documentation (checked 2026-11-07) says it supports both MobileApplication and WebApplication and lists no separate fields for them. Only desktop software and phone apps use SoftwareApplication and MobileApplication respectively.

Which fields does Google require? Which are required and which are recommended?

Only three are required: name, offers.price, and one of aggregateRating or review. Two are recommended: applicationCategory and operatingSystem. The table below is Google's Software app documentation (checked 2026-11-07) combined with the schema.org definitions.

FieldWhat Google requiresHow to write it
nameRequiredThe product name, written the same way as on the page and in the page title
offers.priceRequiredA number as a string; when the price is above 0 also write priceCurrency (for example USD), otherwise Google guesses the currency
aggregateRating or reviewOne of the two, requiredOnly ratings given by real users; if you have none, leave it out (see the next section)
applicationCategoryRecommended, and the value must be one the documentation listsCommon values: BusinessApplication, DeveloperApplication, DesignApplication, EducationalApplication, UtilitiesApplication
operatingSystemRecommendedWrite Web for a web app; otherwise a specific system such as Windows 11, macOS or Android 14
description, url, featureList, screenshot, softwareVersionNot required by the documentationschema.org properties; they help machines read the page, but must match the text people can see on it

Ratings are the strictest part for Google's rich result: by the documentation they are required, so a page without real ratings does not meet the conditions for the software app rich result (this is our inference from the documentation; Google does not state "no rating, no rich result"). Without a rating you can still write the other fields for machines to read.

How do you write the price in offers? Several tiers, a range, a price of 0?

One Offer per tier, the price as a number, the currency written separately. Google's documentation: for an app whose price is 0, write price 0; for a price above 0, include priceCurrency. With several tiers, use an offers array with one Offer per tier, each with a name.

SituationHow it is writtenSources
One priceoffers is a single Offer: price and priceCurrencyGoogle Software app documentation, 2026-11-07
Several tiersoffers is an array with one Offer per tier, each with name, price and priceCurrencyschema.org Offer; Google's field requirements applied to each Offer (our reading)
An app whose price is 0Write price 0Google Software app documentation, 2026-11-07
You only want to give a price rangeAggregateOffer: lowPrice, highPrice, offerCount, priceCurrencyschema.org allows it; Google's documentation lists offers.price and AggregateOffer has no price field, so whether Google accepts it is unverified

Below is a WebApplication example with two price tiers (the name and URL are placeholders; the prices are the two tiers of a kagentx evaluation). The prices must match what the page says.

<script type="application/ld+json">
{
  "@context": "https://schema.org",
  "@type": "WebApplication",
  "name": "ExampleApp",
  "url": "https://example.com/",
  "description": "ExampleApp turns a URL into a one-page report on how AI search sees the site.",
  "applicationCategory": "BusinessApplication",
  "operatingSystem": "Web",
  "offers": [
    { "@type": "Offer", "name": "Basic", "price": "1", "priceCurrency": "USD" },
    { "@type": "Offer", "name": "Full", "price": "19", "priceCurrency": "USD" }
  ]
}
</script>

How to write the price on the page so AI can quote it is in How should a pricing page be written so AI quotes it?.

Can you write ratings and reviews into the markup yourself?

No. Google's Review snippet documentation (checked 2026-11-07) says ratings must come directly from users and may not be created or compiled by editors; if you use AggregateRating, users must be able to see it on the page; and pages of a LocalBusiness or Organization that controls the reviews about itself are not eligible for the star review result. That document has no separate rule for software apps, so the general rules apply, including no fake reviews and no undisclosed incentivized ones.

ApproachAllowed?Cause
Put real ratings and counts from a third-party review site into aggregateRatingYes, provided the page shows that data and names the sourceThe rating comes from users and is visible on the page
Give your own product 4.8 and put it in the markupNoIt is not a rating given by users; that is faking it
The page shows no reviews but the markup has aggregateRatingNoThe markup must match what is visible on the page
No rating yet, so leave out aggregateRating and reviewYesWrite the other fields as usual; for rich result eligibility see the inference in the previous section

What do the kagentx, MochiStamp and Gamma AI Slide home pages actually output?

All three of our sites now have a WebApplication; kagentx added it on 2026-11-07 following this guide, and had only a WebSite before. The table is what curl read from the application/ld+json on each homepage on 2026-11-07 (MochiStamp: the Japanese and English homepages; Gamma AI Slide: the English homepage; in the kagentx column, items marked as added are the state after the fix, the rest is what the English and Simplified Chinese homepages showed before it).

FieldkagentxMochiStampGamma AI Slide
Main type on the home pageWebSite, Organization, WebApplication (added 2026-11-07; before that only WebSite)WebApplication (in one @graph with Organization and FAQPage)Same as left
name, urlYesYes, plus alternateName (モチスタンプ)Yes
descriptionYes (added 2026-11-07)YesYes
applicationCategoryBusinessApplication (added 2026-11-07)DesignApplicationBusinessApplication
operatingSystemWeb (added 2026-11-07)WebWeb
offersTwo Offers: Basic USD 1 and Full USD 19 (added 2026-11-07; the pricing page /pricing/ also got Offers and a BreadcrumbList, and had no JSON-LD before)AggregateOffer: USD 5.99–59.99, 4 tiers, no priceSame as left
aggregateRating, reviewMissingMissingMissing
featureList, screenshot, softwareVersionMissingMissingfeatureList has 7 items; the other two are missing
Organization and FAQPageOrganization yes (added 2026-11-07), FAQPage missingYesYes

The table shows three things. First, kagentx had the biggest gap: the homepage had only WebSite, with no WebApplication or offers, and the pricing page had no JSON-LD at all. On 2026-11-07 we filled it in following this guide: the homepage gained Organization and WebApplication (one Offer per price tier), the pricing page gained Offers and a BreadcrumbList, and there is still no rating. Second, MochiStamp and Gamma AI Slide write their prices as a range with AggregateOffer and no per-tier price, which does not match the offers.price Google's documentation asks for; we have not checked with the Rich Results Test whether Google accepts it, so that is unverified. Third, none of the three sites has a rating, and that is fine: none of the pages has public user reviews, so none should be invented.

Will this markup really make AI more likely to recommend me?

There is no evidence that it will. Google says no extra structured data is needed for its AI features, and the OpenAI and Perplexity crawler documentation do not mention it either (how we checked is in Does structured data help AI search?). The practical benefit of writing WebApplication is that the product name, category, system and price are stated in a way machines can pin down, consistent with the page text, so there is one less chance of being misread. For the state of FAQ markup see Does FAQ structured data help AI search?, for how a SaaS does this overall see How does a SaaS get recommended by AI search?, and for trust pages see E-E-A-T and author pages: do you need them in the age of AI search?.

How do you add it and check it is right, step by step?

  1. 1. Make the page text clear first

    The product name, a one-sentence definition, the price and the supported system are all visible on the page; the markup only restates them.

  2. 2. Choose the type and the category

    For a web app choose WebApplication, pick one applicationCategory from Google's list, and write Web for operatingSystem.

  3. 3. Write offers

    One Offer per tier with name, price and priceCurrency, matching the pricing page; write 0 for a price of 0.

  4. 4. Ratings: only if they are real

    If you have no real ratings from third parties or from users on your site, leave the whole block out; if you do, show them on the page too.

  5. 5. Put it in the home page head

    Use script type="application/ld+json"; put WebApplication on the home page, and the pricing page can carry the same offers.

  6. 6. Check with the Rich Results Test

    Open Google's Rich Results Test, enter the home page URL, and see whether Software app is detected and whether there are errors or warnings. Passing does not guarantee a rich result.

  7. 7. Change the markup whenever the price changes

    When a price changes, change the page, the pricing page and the markup together, and update the sitemap's lastmod.

WordPress.org plugin pages already carry SoftwareApplication markup; for what plugin authors still need to add, see How do WordPress plugins get recommended by AI?.

How WebApplication markup is written on live tool pages and how many impressions each tool page gets is in Do small tool pages (checkers, converters) bring search impressions? Data from MochiStamp's 18 tool pages.

FAQ

Should a SaaS use SoftwareApplication or WebApplication?

For a product used in the browser write WebApplication, a subtype of SoftwareApplication with the same fields. Google's Software app documentation supports both MobileApplication and WebApplication (checked 2026-11-07).

What are SoftwareApplication's required fields?

Google's documentation lists three required: name, offers.price, and one of aggregateRating or review; it recommends applicationCategory and operatingSystem (checked 2026-11-07).

I have no user reviews. Can I write a rating myself?

No. Google requires ratings to come directly from users and to be visible on the page. With no real rating, leave out aggregateRating and write the other fields as usual.

How do I write several price tiers?

Write one Offer per tier in an offers array, with name, price and priceCurrency, matching the prices on the page. AggregateOffer's lowPrice and highPrice are a range that schema.org allows, but Google's documentation lists offers.price, and whether Google accepts a range is unverified.

If I add SoftwareApplication markup, will ChatGPT recommend me?

There is no evidence. Google says no extra structured data is needed for its AI features, and the OpenAI and Perplexity crawler documentation do not mention it. What it does is make the product name, category and price clearer, consistent with the page text.

Sources

kagentx

Can machines read your product page?
Let the evaluation tell you

The evaluation checks whether crawlers can get in, whether the text is in the HTML and whether structured data is present, then asks ChatGPT, Perplexity and Claude (with web search on) your buyers' questions and sees whether they mention you. Basic is $1, usually 2–3 minutes.