Every page is also a text file

What comes back

Open yourname.tapmy.link/page.md — or add the query format=md to the address, or ask for text/markdown in the Accept header — and instead of the rendered page you get a short document: the name, the bio, the template and layout, the language, whether there is a photo, and then every visible block in order with its destination.

It is produced from the same record the templates render, by the same rules. Hidden blocks are absent. Blocks outside their date window are absent. Adult destinations are withheld, exactly as they are withheld from the HTML. If something is missing from the text, it is missing from the page.

1. 01An assistant that just built your page reads it back and compares it with what you asked for.
2. 02A weekly check opens it, walks every URL in it, and reports the dead ones.
3. 03A person can read it too: it is the fastest way to see a long page as a list.
4. 04A page that is unpublished answers 404 here as well, so the text view never leaks a draft.

Why not just parse the HTML

• The HTML is designed for eyes: ten templates, three layouts, a different structure per skin. The text view is one shape forever.
• Parsing markup means re-deriving rules — visibility, date windows, the age gate — that the server already applied.
• A short document costs a fraction of the tokens a rendered page costs, which matters when a model reads it on every check.
• It is stable on purpose. The HTML may be redesigned; the text view is a contract.

What it is not

It is not an export. There are no ids, no timestamps, no settings you cannot see on the page; for that there is the JSON export in your account, or the API with a key. The text view deliberately shows exactly what a visitor can already see, which is why it needs no authentication at all.

And it is not a feed. It has no history and no diff — if you want to know what changed, keep the last copy and compare, which is two lines in any language and the usual way an agent watches a page it maintains.