Why serve Markdown
A modern HTML page is mostly navigation, scripts, styles and tracking. An AI agent that only wants the content has to strip all of that and guess which part matters. A Markdown twin hands it the headings, paragraphs, lists and links directly, in a format models handle well and that uses far fewer tokens.
How to serve one without hurting SEO
The HTML page stays the canonical document. The Markdown twin should send a Link header with rel="canonical" pointing at the HTML URL, and an X-Robots-Tag: noindex, follow header, so search engines never treat it as duplicate content. The HTML page can advertise its twin with <link rel="alternate" type="text/markdown">.
How Aback Launch does it
Any Aback Launch page answers in Markdown when asked with Accept: text/markdown, or when .md is added to the URL. Launch pages have a hand written twin with the listing's description, facts, tags and links. All twins point back to the HTML page as canonical.
Serving it correctly
Answer with Content-Type: text/markdown and a charset. Use the same URL with content negotiation, a parallel URL ending in .md, or both. Keep the twin in sync with the HTML: generate it from the same data rather than maintaining two copies by hand, or the twin will drift and an assistant will quote stale facts.
Strip what does not help a reader, such as navigation, cookie banners and ads, but keep headings, lists, tables, links and the facts a person would want to cite, such as prices and dates.
Example
curl -H "Accept: text/markdown" https://abacklaunch.com/pricing returns the pricing page as Markdown.
Questions people ask
- Will a Markdown twin be indexed as duplicate content?
- Not if it carries a canonical Link header to the HTML page and a noindex robots header.
Related terms
- llms.txtllms.txt is a Markdown file at the root of a website that gives large language models a short summary of the site and links to its most useful pages, so they can read it without parsing every page.
- Generative Engine Optimization (GEO)Generative Engine Optimization (GEO) is the practice of making your content easy for AI assistants such as ChatGPT, Claude, Perplexity and Google's AI Overviews to find, understand and cite in their answers.
- Canonical URLA canonical URL is the one address you want search engines to index for a page that can be reached at several. You declare it with <link rel="canonical" href="..."> in the page's head.
- NoindexNoindex is a directive that asks search engines to keep a page out of their results. It is set with <meta name="robots" content="noindex"> or an X-Robots-Tag HTTP header.