iTechGuides is reader-supported. When you buy through links on our site, we may earn an affiliate commission. As an Amazon Associate I earn from qualifying purchases. Learn more
Yes: a website can return a Markdown version of a page when a client requests it with the HTTP header Accept: text/markdown. This is content negotiation: the same page URL can provide HTML to ordinary browsers and Markdown to clients that ask for it. Markdown can give software a cleaner, more predictable representation to process, but it does not guarantee that agents will request it or improve the accuracy of their answers.
How Markdown for Agents works
A client requests a page with Accept: text/markdown. If the site or its delivery platform supports the format, it converts the page and returns Markdown instead of HTML. The request uses the ordinary page URL; the header asks for an alternate representation.
| # | Preview | Product | Price | |
|---|---|---|---|---|
| 1 |
|
The Markdown Guide | $7.95 | Buy on Amazon |
| 2 |
|
Using Markdown: A Short Instruction Guide | $9.99 | Buy on Amazon |
| 3 |
|
Markdown: A Complete Guide | $9.99 | Buy on Amazon |
| 4 |
|
Accessible Markdown: Structured Authoring and Reliable Exports | $19.99 | Buy on Amazon |
| 5 |
|
R Markdown Cookbook (Chapman & Hall/CRC The R Series) | $25.31 | Buy on Amazon |
For example, Cloudflare documents conversion at its network edge for enabled zones: it fetches the origin’s HTML and converts it before returning a response when possible. The site’s origin can remain the source of the page. Read the Docs documents a similar behavior for its hosted documentation domains, where Markdown delivery is enabled automatically and ordinary browsers continue to receive HTML. Cloudflare’s Markdown for Agents documentation and Read the Docs’ Markdown documentation describe these implementations.
This approach can spare a consuming client from independently extracting and simplifying content from rendered HTML. It is an implementation goal, not evidence of better model accuracy, answer quality, search ranking, or universal token savings.
#1 Best Overall
What the request and response look like
A minimal request has the same path a browser would use, with an additional Accept header:
GET /fundamentals/reference/markdown-for-agents/ HTTP/2
Host: developers.cloudflare.com
Accept: text/markdownCloudflare’s documented converted response uses Content-Type: text/markdown; charset=utf-8 and adds Accept to Vary. That variation matters: caches need to distinguish a Markdown response from an HTML response for the same URL. Cloudflare also documents recalculating Content-Length, adding estimated x-markdown-tokens and x-original-tokens headers, preserving security- and cache-relevant origin headers, and removing body-specific headers such as ETag and Last-Modified because conditional requests cannot be honored for the converted body. These are Cloudflare’s documented behaviors, not guarantees for every implementation. See its response-header details.
The documented Cloudflare body can contain optional YAML frontmatter generated from supported metadata tags, converted page content, and a fenced JSON block containing source JSON-LD when present. Its preprocessing removes non-content elements such as navigation, scripts, and styles. The resulting Markdown may therefore include metadata and structured data in addition to the main text.
Recommended Free Tools
How to enable Markdown delivery
Cloudflare: convert pages at the network edge
Cloudflare’s documentation describes a zone-level setting, with configuration rules available to limit the feature to selected hostnames or paths, as well as an API route for changing the zone’s content_converter setting. The documented dashboard route requires a Pro or Business plan for the zone setting. Because plan eligibility, labels, and API behavior can change, check the current Cloudflare setup documentation before enabling it.
Rank #3
Pay attention to the response’s content policy signal. Cloudflare documents that, when the origin does not send a Content-Signal header, it adds Content-Signal: ai-train=yes, search=yes, ai-input=yes by default. If the origin supplies a value, Cloudflare preserves that value as authoritative. Review and set the origin’s preferred policy before switching on conversion. The header behavior alone does not establish what legal effect, if any, a signal has.
Read the Docs: use hosted documentation delivery
Read the Docs says Markdown responses are automatically enabled on its hosted documentation domains. This is a feature of that hosting platform, not a setting that applies to arbitrary websites. If your documentation is hosted there, its official guide explains the behavior.
Other implementation paths
The same outcome can be built in an origin or application pipeline, or through other infrastructure that supports content negotiation. The reviewed provider documentation establishes examples rather than an exhaustive list of options. When evaluating an approach, check where conversion runs, which pages it covers, how it handles caches and metadata, and whether it fits your existing hosting and content-update workflow.
Markdown for Agents, llms.txt, and agents.txt are different
These conventions and mechanisms address separate problems. They can coexist, but one does not replace the others.
Best Value
| Mechanism | What it does | What it does not do |
|---|---|---|
| Markdown for Agents | Provides an alternate representation of a requested page, typically through Accept: text/markdown. |
Does not itself advertise every useful page or declare what actions an agent can perform. |
llms.txt |
A proposed, non-ratified convention for a curated index of important site content. The explainer describes a v2 discovery approach using a rel="describedby" link and path-scoped indexes. |
Does not grant or deny access, protect omitted pages, or establish that model vendors’ crawlers will read it. |
agents.txt |
A capability-declaration layer describing protocols or endpoints a site supports. | Does not provide the page’s content representation or act as a curated content index. |
The llms.txt explainer, dated 2026-08-21, characterizes the convention’s status and notes that vendors do not disclose whether their crawlers read the file. The agents.txt project README distinguishes content briefing from capability declaration. Access and use policies are handled separately from an llms.txt index.
Quick Recap
What Markdown delivery does—and does not—establish
- It offers a requestable format: clients that support the mechanism can ask for Markdown while browsers keep using HTML.
- It can simplify content processing: the converted representation may remove page elements that are not content and retain useful metadata or structured data.
- It is not universal: a site must implement the behavior, and a client must request and use the response. The reviewed sources do not establish broad agent uptake.
- It is not a measured performance promise: Cloudflare’s token-count headers are estimates, and sample response values are illustrative—not a general study of savings.
- It does not make a site fully agent-ready: content representation, discovery, access policy, and action capabilities are distinct concerns.
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
