Frequently Asked Questions

AIPROPX is a deterministic, multi-source index of real-world news events. Right now the ledger holds 260 resolved story clusters drawn from 482 tracked outlets. Each cluster is one permanent canonical entry per event — built by computing text overlap between publisher reports. We index and resolve; we never republish original article content.

AIPROPX is an independent one-person operation. There is no newsroom, no editorial team, no committee. One person built it and one person runs it. That constraint is a feature, not a limitation: a system with a single operator cannot have editorial drift, hidden biases introduced by staff rotation, or institutional pressures to favour any source. Deterministic by necessity, and by design.

No. AIPROPX is not a publisher. We do not write headlines, publish articles, or commission journalism. We read public news feeds, compute text overlap between reports from different outlets, and group them into canonical entries. Every source is named and linked at origin. Opening a report sends you directly to the original publisher; AIPROPX holds no copy of their article.

It means every figure you see is computed, not decided. Outlet counts, report counts, publication timestamps, and text-overlap scores are all the output of algorithms running against real feed data. No human writes a summary, no AI interprets the news, no editorial process assigns weight to any source. The same inputs always produce the same outputs. That is what "deterministic" means in practice.

Our system computes text overlap between reports from different publishers. When multiple outlets cover the same underlying event, their reports are grouped into a single canonical entry. No editorial judgement is applied — it is a purely algorithmic process. A story with a high outlet count is not "confirmed"; it means many newsrooms published their own version of the same event.

When a story entry shows an outlet count, that is the number of distinct publishers whose reports were grouped into that cluster. It is a count, not a credibility score. A count of 1 means one outlet reported it; a count of 10 means ten different newsrooms each wrote their own version. We currently track 482 outlets across every region.

No. AIPROPX never rates, scores, or labels any outlet, report, or claim as accurate, biased, politically leaning, or credible. When we display that a sentence appears in multiple reports, that is a count of text matches — not a verdict on whether it is correct or trustworthy.

"Region" reflects only where the publishing outlet's headquarters is registered. It is never an inference about where the event occurred, who it concerns, or the political stance of the coverage. A report filed by a US-based outlet and a report filed by a UK-based outlet about the same event will each carry their outlet's region — not the event's location.

The ingestion pipeline runs continuously. New reports are pulled from publisher feeds and clustered on a short cycle throughout the day. Story clusters update as new outlets cover an event — you will see the outlet count rise and the timeline extend as coverage develops.

Stories come from real publisher RSS and Atom feeds, scraped on a scheduled cycle. Every article in a cluster links directly back to the original publisher. AIPROPX does not generate, summarise, or paraphrase any article text for display purposes.

Yes. Our clustering algorithm, scoring formulae, and data sources are documented on the Method page. Readers can verify the counts independently by opening any story entry and clicking through to the original publisher articles. We publish methodology revisions with version notes on the Editorial Standards page.

The Smart Ledger is an on-demand synthesis of a story cluster — a structured summary produced by a language model on explicit user request. It is clearly labelled as AI-generated and is never served automatically on page load. No AI touches the main feed, the outlet counts, or any other index figure. The Ledger page has a live demo where you can inspect a real cluster without any AI involvement.

Visit the Corrections page or email [email protected]. Include the URL of the affected entry, a clear description of the error, and any source links that support your correction. Anonymous submissions are accepted. We review every request within 48 hours.

We correct the record immediately and publish a dated correction note on the affected entry. We do not silently amend indexed data. The original error text is preserved in the note for full transparency. A public log of confirmed corrections is maintained on the Corrections page.

No. We can only correct errors in our own indexing — wrong outlet attribution, wrong timestamp, wrong cluster assignment, and similar. If an original article contains a factual error, that is between the reader and the publishing outlet. We do not edit, summarise, or modify any publisher's content.

Yes. Reading the ledger, browsing topics, following sources, and viewing trending stories is completely free. An account lets you personalise your feed, save topics, and manage your newsletter preferences.

Click "Account" in the top navigation and choose "Sign in". You can register with your email address or connect a social provider such as Google or GitHub. No credit card is ever required.

Sign in, go to Account then Communications, and toggle each email type on or off. Changes are saved instantly. You can also unsubscribe from any email we send using the unsubscribe link at the bottom of each message.

Sign in and go to Account then Privacy and Data then Delete account. Enter "DELETE" to confirm. This permanently removes your account, sign-in methods, and email subscriptions and cannot be undone.