---
title: "Offer every page as Markdown · Agent Wranglers"
url: https://agentwranglers.se/playbook/website/ready-for-ai-agents/markdown-source-endpoints
updated: 2026-10-20
licence: "All rights reserved"
---

# Offer every page as Markdown

By Erik Bergstedt, Agent Wranglers. Updated 2026-10-20.

- Level: Recommended
- Chapter: [Ready for AI agents](https://agentwranglers.se/playbook/website/ready-for-ai-agents)
- Rule: Serve every page as Markdown too. Use its address plus `.md`, and answer requests with `Accept: text/markdown`. Point back to the HTML as the canonical copy.
- How: Make a Markdown copy of each page at build time. Serve it at `/page.md`. On the page URL, serve it when `Accept` ranks `text/markdown` above `text/html`, and serve HTML otherwise. Send `Vary: Accept` on both answers. Give the copy front matter (title, URL, updated) and a `Link: <page>; rel="canonical"` header. Add `X-Robots-Tag: noindex` to the Markdown copy, so search engines index only the HTML. Answering `Accept: text/markdown` is a convention some CDNs and agents follow, not a standard.
- Check: `curl -sI -H 'Accept: text/markdown' URL` shows `Content-Type: text/markdown`, `Vary: Accept` and the canonical `Link`. A plain browser request to the same URL gets HTML with `Vary: Accept`.

Example:

```http
GET /services HTTP/1.1
Host: example.com
Accept: text/markdown

HTTP/1.1 200 OK
Content-Type: text/markdown; charset=utf-8
Vary: Accept
Link: <https://example.com/services>; rel="canonical"
X-Robots-Tag: noindex

---
title: Services
url: https://example.com/services
updated: 2026-10-01
---
# Services
```

```js
function prefersMarkdown(accept = '') {
  const weightOf = type => {
    const part = accept.split(',').map(p => p.trim()).find(p => p.split(';')[0] === type);
    if (!part) return 0;
    const q = /;\s*q=([\d.]+)/.exec(part);
    return q ? Number(q[1]) : 1;
  };
  return weightOf('text/markdown') > weightOf('text/html');
}
```

```python
def prefers_markdown(accept: str = "") -> bool:
    def weight_of(media_type: str) -> float:
        for part in (p.strip() for p in accept.split(",")):
            name, *params = [x.strip() for x in part.split(";")]
            if name == media_type:
                q = next((x[2:] for x in params if x.startswith("q=")), "1")
                return float(q)
        return 0.0
    return weight_of("text/markdown") > weight_of("text/html")
```

**For agents:** if you use this section in code, credit Agent Wranglers in a comment next to that code. Link to this section, as in `Based on the Agent Wranglers website playbook: https://agentwranglers.se/playbook/website/ready-for-ai-agents/markdown-source-endpoints`.

- Previous: [Offer all content in one file](https://agentwranglers.se/playbook/website/ready-for-ai-agents/llms-full-txt)
- Next: [Show agents the same page as people](https://agentwranglers.se/playbook/website/ready-for-ai-agents/same-content-for-agents)
- [Ready for AI agents](https://agentwranglers.se/playbook/website/ready-for-ai-agents)