What Makes Veridi Different

It has a checklist, and it follows it

Most fact-checking relies on the judgment of experienced people. That works well, but it’s hard to audit, hard to scale, and hard to explain. Why did two fact-checkers reach different conclusions about the same claim? Without a documented process, it’s hard to say.

Veridi uses explicit procedures. Every step, from classifying a claim to choosing a verdict, follows documented rules. When the assessment is done, you can trace exactly how it got there.

It knows what tricks to look for

Disinformation isn’t random. People who spread false information use specific techniques, and those techniques have names and solid definitions.

Here are a few examples:

Confidence laundering. A claim appears on a blog. A news site picks it up. Another news site cites the first one. Now it looks like multiple independent sources confirm the claim - but they all trace back to one unreliable post. Veridi traces every source back to its origin.

Anchoring. A true, easily verified fact is placed next to a false claim in the same sentence. The true part makes the whole thing feel credible. Veridi separates multi-part claims and evaluates each piece independently.

Framing. Every individual fact in a claim checks out, but they’re arranged to create a false impression. This is one of the hardest tricks to catch because nothing is technically wrong; the lie is in the framing, not the facts. Veridi checks whether true facts are being assembled to mislead.

The methodology checks for eleven of these techniques on every assessment. When it finds one, it flags it and explains what’s happening.

It doesn’t trust sources blindly

Not all sources are equal. A government database and a social media post are not the same quality of evidence. Veridi ranks every source on a four-tier scale, from primary sources (uncompromised government data, peer-reviewed research) down to social media and anonymous sources.

Here’s the key: the quality of your sources limits how confident the assessment can be. If the best evidence available is news articles, confidence is capped at 65%. If you have government data and peer-reviewed research in agreement, confidence can go up to 95%. This prevents the system from being tricked by quantity. A hundred blog posts don’t outweigh one verified dataset.

It knows that institutions can change

A government agency that was a reliable source last year might not be this year, if it’s been defunded, had its scientists fired, or had its leadership replaced for political reasons. Veridi tracks this through an Institutional Reliability Index that assesses specific agencies on specific topics, and provides alternative sources to consult when a primary source has been compromised.

It tells you when it doesn’t know

When the evidence is genuinely unclear, Veridi says “unverifiable” rather than guessing. When a topic is legitimately contested among experts, the assessment reflects that disagreement as accurately as possible. The system has nine verdict categories specifically because the world is more complicated than true or false.