Guide · Digital products

How do digital products sold on Gumroad and Lemon Squeezy get recommended by AI?

Updated 2026-11-09 · kagentx

Short answer

Make your own site's product page the main page, let the platform page handle checkout only, and keep price and description the same on both. Tested 2026-11-09: Gumroad's robots.txt blocks only /purchases/, its product page has Product structured data, its canonical points to the platform itself, and the full description is not in the visible HTML markup; Lemon Squeezy's main site blocks only /backups/; Ko-fi disallows training crawlers such as GPTBot and ClaudeBot; our requests to Payhip and Ko-fi pages were stopped by a verification page, so their content is unverified. Whether AI cites these platforms' product pages is not stated officially, and we have not verified it.

What do these platforms' robots.txt files say to AI crawlers?

They differ a lot, and the platform sets them all; you cannot change them. On 2026-11-09 we read the robots.txt of four platforms: Gumroad and Lemon Squeezy barely restrict anything, Payhip blocks only checkout and download paths, and Ko-fi explicitly disallows training crawlers such as GPTBot and ClaudeBot. None of the four names OAI-SearchBot, ChatGPT-User or PerplexityBot.

Platformrobots.txt (2026-11-09)What it tells you
gumroad.comUser-agent * blocks only /purchases/; about 374 Sitemap lines; no AI crawler namedProduct pages are not blocked
lemonsqueezy.comThe main site blocks only /backups/; a store subdomain's robots.txt returns 404 (no rules)The main site's rules are permissive; merchants' storefront pages are unverified (see below)
payhip.comBlocks /d/, /e/, /s/, /download/, /buy, /order, /checkout, /review and so on; has a Sitemap index; no AI crawler namedCheckout and download pages cannot be crawled; store and product pages are not blocked
ko-fi.comContent-Signal ai-train=no; Disallow / for GPTBot, ClaudeBot, anthropic-ai, CCBot, Google-Extended and others; OAI-SearchBot and PerplexityBot not namedTraining crawlers are blocked; search crawlers are not restricted by the rules

Rules are not the same as what can actually be fetched: for Payhip and Ko-fi, curl got an HTTP 403 verification page on the home and shop pages, so their page content is unverified (could not connect).

Can AI read product pages on these platforms?

We only got a sample from Gumroad. Gumroad product pages are not blocked and carry Product structured data, but the full product description is not in the visible HTML markup. We could not get product pages from Lemon Squeezy, Payhip or Ko-fi, so those are unverified.

What we found (Gumroad, 2026-11-09)ResultWhat it tells you
titleThe product page has only the product name (CodegridPRO); the store page has the store nameThe title does not say what the product is, so the product name has to carry its own meaning
meta descriptionThe first roughly 150 characters of the description, plus an ellipsisThe first paragraph of the description is the public summary
Structured dataProduct: name, description, brand, offers, aggregateRatingPrice and rating are given as structured data; its description is not the same text as the meta description
canonicalPoints to the product's own URL on the Gumroad subdomainA platform page's canonical points to the platform itself, not to your own site
Page bodyThe visible HTML has no h1 and no body paragraphs; the full description sits in embedded JSON (the data-page attribute)Whether AI crawlers parse this is unverified

A Lemon Squeezy storefront is off by default (official help docs, viewed 2026-11-09); you can turn it on, or set a custom redirect so that visits to the store address go to your own site. A custom domain is used in product share URLs and checkout URLs. So on Lemon Squeezy, you decide which domain the product page lives on.

Which should be the main page: the platform's product page or your own site?

We suggest making your own site the main page and letting the platform page handle checkout only. The reason: you cannot change the platform page's title, markup, robots.txt or canonical, and its canonical points to the platform itself, while you control every part of your own page. This is our recommendation, not a rule from the platforms or search engines.

LocationWhat goes inWhy
Product page on your siteDefinition, who it is for, what is included, price, update date, FAQ, comparison; the buy button links to the platform's checkoutYou control the title, structured data and canonical
Product page on the platformThe same price and contents; a first paragraph that means the same as on your site; a link back to your siteHandles the purchase; the description and price must not contradict your site
Brand nameUse the same product name on bothAI uses the name to recognise the two pages as the same product

The platform page and your own page each exist independently, and AI may read either, so update price, contents and version number on both at the same time. For the general way to set canonical, see Canonical tags and duplicate content: how to set them up on a multilingual site so pages are not treated as duplicates.

How do you write the first paragraph of a product description?

Treat the first paragraph as an answer that can be quoted directly: what the product is, who it is for, what it includes, how it is delivered. Gumroad's meta description takes the first roughly 150 characters of the description, so the most important information has to fall inside that.

  1. 1. Start with the definition

    “This is a 120-page Notion finance template for freelancers to manage invoices and taxes.” Do not write “powerful” or “ultimate”.

  2. 2. Follow with what is included

    File format, page or item count, how updates work, for example “a PDF plus a Notion template link, lifetime updates”.

  3. 3. Then who it is and is not for

    One sentence on the prerequisites, such as needing a Notion account.

  4. 4. Write price and version as text

    Price, license scope (personal or commercial) and last update date; do not leave them only in images.

  5. 5. Keep both sides in sync

    The first paragraphs of your site's product page and the platform's product page should mean the same; the wording can differ.

The names and numbers in the example are placeholders.

Where do comparison and alternatives pages belong?

Put them on your own site, not on the platform page. You cannot change a platform page's layout or markup, and it has no place for a comparison table; your site can carry a table with sources and dates that answers “what is the difference between X and Y”. For how to write them, see How do you write an “X vs Y” comparison page that AI cites?; for alternatives pages, see How to write an “X alternatives” page that AI will cite.

The platforms do this themselves: Lemon Squeezy's sitemap lists /gumroad-alternative and /paddle-alternative (we only read the addresses and did not analyse the content).

Your site also needs a page that says who you are, see How to write an About page so AI gets your company and product right; if you also sell physical goods, see How do you get ChatGPT to recommend your product on an export-focused independent store?; first use AI crawler check to confirm your site does not block AI crawlers, and see How do you write robots.txt for AI crawlers? for the rules.

If you sell Notion templates, Figma Community files or Framer templates, the robots.txt and raw-HTML tests of the three marketplaces are in How do you get Notion templates, Figma Community files and Framer templates recommended by AI?.

Real data on digital products like sticker and emoji packs in search and AI answers is in How do sticker and emoji packs get found in search and AI answers?.

If you build a Blender add-on, see How do AI search tools find Blender add-ons? The Extensions platform tested, plus the official listing requirements for the Extensions platform's robots.txt, the detail page HTML test and the official listing requirements.

If you make Unity assets or tools, see How do AI search tools find assets on the Unity Asset Store? Asset pages tested, plus the official rules for the Asset Store's robots.txt, sitemap, asset page test and the official Submission Guidelines.

FAQ

Will ChatGPT cite a Gumroad product page?

We found no official statement and no evidence of actual citations, so this is unverified. What we can confirm: robots.txt blocks only /purchases/, and the product page has Product structured data (tested 2026-11-09, one product page as the sample).

Where should the canonical point, the platform page or my site?

The platform generates a platform page's canonical; the Gumroad product page we saw points to itself, and you cannot change it. So we suggest making your site the main page, keeping the platform page's price and description consistent with it, and linking back to your site. This is our recommendation, not a platform rule.

How long should the first paragraph of a product description be?

Make sure the most important information falls within the first 150 characters: the meta description of a Gumroad product page takes roughly the first 150 characters of the description plus an ellipsis (tested 2026-11-09). We have not verified the cut-off length on other platforms.

Ko-fi blocks GPTBot. Does that mean ChatGPT will not recommend me?

Not necessarily, but we have no evidence either way. Ko-fi's robots.txt disallows training crawlers such as GPTBot and ClaudeBot and does not name OAI-SearchBot or ChatGPT-User. Rules only govern whether a crawler may fetch a page; whether an AI answer cites it is not stated officially and is unverified.

Can comparison and alternatives pages live on Gumroad or Lemon Squeezy?

We do not recommend it. You do not control a platform page's layout or markup, and there is limited room for a comparison table; on your own site you can state sources and update dates.

Sources

kagentx

Can crawlers read your pages?
Let the evaluation tell you

The evaluation checks whether crawlers can get in and whether the text is served directly, then asks ChatGPT, Perplexity and Claude (with web search on) your buyers' questions and sees whether you are mentioned. The Basic plan is $1, and results usually take 2–3 minutes.