fetch-as-markdown

Files (1)

• SKILL.md

SKILL.md (first lines)

# Fetch juanlentino.com content as Markdown

Any page on this site will answer in Markdown if you ask for it. Use this when
you want the prose without parsing HTML.

## How

Send `Accept: text/markdown` with an ordinary GET:

```
GET https://juanlentino.com/notes/ HTTP/1.1
Accept: text/markdown
```

The response is `text/markdown`. Without the header you get the normal HTML,
and the URL does not change.

A note also has its Markdown at its own URL, for clients that cannot set a
header: append `.md` to the note URL (`/notes/<slug>/` becomes
`/notes/<slug>.md`). It is the same Markdown, and each note page links it as
`<link rel="alternate" type="text/markdown">`. Other pages answer `.md` with
a 404; use the header there.

## Notes

- Conversion happens at the edge, after the cache lookup, so `Vary: Accept`
  rides both representations and a cached Markdown body cannot reach a browser.
- Measured on a live note: 131,594 bytes of HTML became 6,643 of Markdown.
- Structured alternatives exist if you want data rather than prose: every note
  has a `.json` twin, and `/feed/json/` lists them.

## Validate

```
curl -s -H 'Accept: text/markdown' https://juanlentino.com/notes/ | head -5
```

You should see Markdown, not `<!DOCTYPE html>`.

From juanlentino/signal-and-noise-tools, GPL-2.0.

More from juanlentino/signal-and-noise-tools