Avyth

For newsletter writers

Own the archive, and disclose what you should.

Two tools for people who publish on someone else's platform: one keeps a copy of everything you have written in formats that outlast the platform, and one checks whether a site that uses AI says so in the way EU law now asks.

The week, in order

  1. 1 · Backing up the publication

    Substack Archiver

    What it does →

    Takes a name.substack.com address and returns the public posts as HTML and Markdown with a linked index — an archive you own, readable without Substack. Avyth's server does the fetching, so this is the one tool here that is not browser-local.

    Up to 200 public posts as HTML and Markdown, free; the subscription bundles the images so it reads offline, adds an EPUB of the whole publication, and reaches 2,000 posts.

  2. 2 · Checking the site's AI disclosures

    Article 50 Check

    What it does →

    Reads the page your server delivers and lists the EU AI Act Article 50 gaps — where AI-generated text, images or a chatbot should be disclosed and is not — each with wording that would close it.

    A full findings report with suggested wording for one page, free; the subscription scans multiple pages, gives the wording in all 24 EU languages and produces an evidence PDF.

What it costs

Substack Archiver is $5 a month or in the $8 Creator bundle; Article 50 Check is $5 or in the $12 Documents bundle; every paid tool is $24. The free tiers cost nothing and need no account.

What this page won't pretend

  • Subscriber-only posts. The archiver fetches public posts with no login, so paywalled content is listed as skipped, not copied.
  • Custom domains. It needs the original name.substack.com address today.
  • Legal advice. Article 50 Check is a review with suggested wording, not a certificate of compliance, and it cannot see a chat widget injected by JavaScript after the page loads.

Questions

Does the archiver bypass paywalls?
No. It requests posts the way a logged-out reader would; anything not public is skipped and listed as such in the archive.
Why is the archiver not browser-local like the others?
Because a browser cannot fetch another site's pages directly. Avyth's server fetches the posts, zips them and returns the ZIP; it keeps none of it, and its privacy page says exactly what is logged.
Does Article 50 apply to a newsletter?
It applies to sites and services that deploy AI in the ways the Article names — generated text presented to the public, synthetic images, chatbots. If a newsletter's site does none of those, the check will say so.
Can I get the EPUB without the subscription?
No — the EPUB and bundled images are the paid step. The free archive is complete as HTML and Markdown, with images left at their original addresses.

Other kinds of workCompared with what you might use instead