How TruthRanking works

One workflow, from the first mention of a claim to tracking whether search results still repeat it.

1. Monitor a topic

You enter a topic, the organizations and projects involved, and the phrases people use. TruthRanking watches web pages, news, guest columns, podcasts, and search results, including AI-generated summaries, for new mentions. Posts on social platforms are collected logged-out only, and any post can be added by pasting its link.

2. Isolate each claim

A single article can carry a dozen separate claims. TruthRanking pulls each one out as its own record, quoted word for word and tied to the exact page it came from. Mixed statements are split, so a checkable fact and an opinion in the same sentence are handled separately.

3. Trace it to the first source

Most claims you see are copies. TruthRanking links every repetition back to the earliest publisher it can find, and records who actually said it. A hundred reposts of one press release count as one origin, not a hundred confirmations. Read more in how to find where a claim started.

4. Rate the claim against evidence

Each claim gets one of eight verdicts, from supported to contradicted, plus an evidence level and a confidence level. A verdict that a claim is wrong requires linked contradicting evidence. An unchecked claim is labeled unverified, never false. A second person must approve every verdict before it counts.

5. Prepare the correction

When a claim does not hold up, TruthRanking assembles the case: the claim, its first source, the evidence, and a private, time-stamped capture of the page. You choose the route: an editorial correction request, a public explanatory reply, an owned-site change package, a platform policy report, a search and AI feedback package, an outdated result refresh, or a counsel review packet. Every outbound message is a draft until a person approves it. Nothing sends on its own.

6. Track the result

After a correction goes out, TruthRanking keeps checking the source page and the search results for your topic, so you can see whether the page changed, the post came down, or the summary stopped repeating it.

What it will not do

• It rates claims, never people. There are no verdicts on anyone's honesty, motives, or character.
• It does not build personal profiles. Expert credentials are limited to specialty and a public registry link.
• It does not post on its own, create fake supporters, or flood any platform.
• It does not log in to anyone's personal accounts to collect data.