FAQ

Can I hotlink files?

Yes. That is what the site is for.

CORS is open to every origin. Files carry Cross-Origin-Resource-Policy: cross-origin. Byte ranges work. So an <img>, <video> or fetch() on any page can use a file directly.

Is there a rate limit?

Yes. Keep automated traffic under 30 requests per second per IP. A faster client may be rate-limited.

If a CI job needs many files, download them once and cache them.

Why is there no ?download parameter?

A query string never changes a response. Every request for a path gets the same cached object, so there is nothing for a ?download parameter to switch on.

To save a file, use curl -O, wget or your browser's save command.

Why do search engines not index the files?

Because we tell them not to. Files are served with X-Robots-Tag: noindex. Search results lead to the format pages instead. Those pages explain what each file is and give you its URL.

Can I mirror the files?

Yes. Every file is CC0 1.0, so you can copy it for any purpose without asking.

Check your copy against sha256sums.txt. And please download once rather than over and over.

Will a file ever change?

No. The bytes at a published path never change. If a file turns out to be defective, we publish a replacement at a new path, and the old entry points to it.

In rare cases a file can be withdrawn. The terms of use cover that.

Where do the files come from?

We generate every file with the open-source tooling in the repository. It runs inside a pinned container and starts from synthetic data.

No file contains third-party media or real personal data. The one exception is that some images and PDFs rasterise glyphs from open-licensed fonts. THIRD_PARTY.md in the repository lists them.

Accessibility

The site targets WCAG 2.1 AA; report issues on GitHub.