DISC-009Find
Pages served as Markdown on request
Checks whether your home page returns a real Markdown version of itself when a client asks for Markdown instead of HTML.
What we check
We check whether your home page supports Markdown content negotiation: when a request carries "Accept: text/markdown", the server answers with Content-Type text/markdown and a body that is actually Markdown, while ordinary visitors keep getting HTML. This follows Cloudflare's "Markdown for Agents" (docs page updated 2026-07-13), a vendor convention built on HTTP content negotiation and the text/markdown media type (RFC 7763). It is not a formal standard.
Why it matters
An AI shopping agent reading your pages has to pull the words out of HTML full of scripts, styles and layout markup. A Markdown version gives it the same content in a smaller, cleaner form, so it is cheaper and more reliable for the agent to read what you sell.
How we check
We make one request for your home page with "Accept: text/markdown". PASS: HTTP 200, Content-Type text/markdown and a body that is Markdown. Leading HTML comments are ignored, and a README-style opening line of link or image badges followed by Markdown still counts. WARN: the response is labelled text/markdown but the body is the HTML page (Shopify storefronts currently do this) or is empty. FAIL: the ordinary HTML page or another media type comes back, or the server answers 406, 404 or 410. SKIPPED: bot management answered instead of your shop, your robots.txt disallows the path for our crawler, or the server returned another status such as 429 or a 5xx.
How to fix it
- Answer requests carrying "Accept: text/markdown" with a Markdown version of the page, and keep HTML for every other request.
- If your site is behind Cloudflare (Pro plan and above), turn on Markdown for Agents. Otherwise, convert the page at your CDN edge or in your theme or application.
- Send "Content-Type: text/markdown; charset=utf-8" and "Vary: Accept" so caches keep the two versions apart.
- On Shopify: the storefront currently labels the HTML page as text/markdown without converting it. Converting it needs a layer in front of the store, such as a CDN, until the platform changes.
What good looks like
GET / HTTP/1.1 Accept: text/markdown HTTP/1.1 200 OK Content-Type: text/markdown; charset=utf-8 Vary: Accept # Example Shop Hand-made ceramics, shipped from Leeds. ## Bestsellers - [Stoneware mug](https://example.com/products/stoneware-mug) - GBP 24 - [Serving bowl](https://example.com/products/serving-bowl) - GBP 58
Sources
Read more
- How an AI agent reads a web page
Raw HTML, the rendered page, the accessibility tree, structured data, Markdown versions and llms.txt, and what each one means for an agent trying to read your product page.
Could an agent buy from your store?
The free scan runs all 51 checks on your store in about five minutes.
Scan my store