Privacy

Draft. Published ahead of legal review; the final version follows that review. Even so, it describes what this site actually does.

The controller. Macskásy Attila, registered Hungarian sole entrepreneur (trade name: ATAILA) · registration no. 60271420 · tax no. 90917792-1-41. Contact details are in the imprint. You may request access to your data, rectification, erasure, restriction of processing, object to processing based on legitimate interest, and request data portability (GDPR Arts. 15–21). We respond without undue delay and within one month at the latest. You may lodge a complaint with the Hungarian supervisory authority (NAIH, naih.hu) or turn to the courts.

This website

The product

Two sentences the product rests on, and both are on our change-control list:

Your file, your drafts and your spike log stay with you. Nothing publishes without a named person’s approval.

The first sentence ends the moment the file, the drafts or the spike log are sent to a hosted AI provider or to a third party. The second ends the day anything publishes on its own. If either becomes untrue, this page changes first.

Where your file lives, per edition.

In both editions no hosted AI provider holds your file, your drafts or your spike log.

What leaves. PrivateWire fetches public web pages as part of research: the article, what it cites, what contradicts it. That is a genuine outbound flow, and we draw it ourselves rather than leave you to find it. What stays with you is the file, the prompts, the drafts, the spike log and the decisions.

Who works with your file today. Today ATAILA operates the desk with you: we run the wire and draft; your named person approves. That means named ATAILA staff work with your file, under your contract and under the confidentiality it sets. The desk you log into yourself is on the roadmap, and when it ships this paragraph changes.

What is recorded. Every approval and every desk decision is recorded with who and when, on the piece. Administrative sessions are logged. The record is yours, available on request, and it is part of the product, not a by-product of it.

Three different things. The file is your profile, your positions and your prior work. The spike log is what you declined and why. Both are yours. The archive is our captured copy of somebody else’s page, and it is governed by the next section.

The archive

  • What is captured. Every article we answer is captured in several forms, hashed, dated, and kept beside the answer. For ATAILA’s own newsroom every capture is dated 6 September 2026; the archive proves what each page said on that day.
  • What is public. While the original is alive, the public sees provenance: the publisher, the headline, the date, the capture time, the hashes, a thumbnail, a short extract, and a live link that passes authority to the publisher. The faithful copy of the page is never public while the source is alive, and the content schema refuses to represent a public extract while the source is alive.
  • What is private. The full copies sit behind a sign-in, as a record. A captured page may contain the personal data of its author or of people quoted in it; that data is held as part of the private record and is not published by us.
  • A publisher who asks is moved to provenance-only, without argument: headline, date, link, nothing else. Write to the address in the imprint.
  • Retention is a contract term set with counsel. No period, no tier and no default retention is promised on this site.
  • Archive pages are not indexed and are not listed in our sitemap. That is the difference between preserving a record and republishing someone’s work into search results.

AI transparency

Changes

This page changes before the product does. If a hosted AI provider ever enters the loop for a customer’s file, or if anything ever publishes without a person, this page is corrected first and the footer line on every page comes down the same day.