How we review
What our evidence labels mean, where our facts come from and what happens when we get something wrong.
The three evidence levels
Every article states, near the top, how much first-hand evidence sits behind it. We pick the label honestly and change it if the evidence changes.
Documentation-based review
Built from the product's official documentation, pricing and policy pages, not from our own hands-on use. These articles explain what a product says it offers, what it costs and what its policies say. They do not claim personal experience of using it.
Hands-on tested
Our editors used the product directly for this piece, alongside its official documentation. The article will say what we did, on what plan and when. We only use this label when an editor has actually used the product for the piece.
Not yet independently tested
We have not verified these claims ourselves yet. Treat product details as reported, not confirmed. We use this label for early coverage, news about a product, or where official documentation is thin.
How we source facts
- Primary sources first: official pricing pages, terms, privacy policies, help centres and app store listings.
- Dated: every source is listed at the end of the article with the date we accessed it, because AI prices and features change often.
- No invented numbers: we do not make up statistics, user counts, test scores or star ratings. If we do not know something, we say so.
- Clear wording: when something comes from the company, we present it as the company's claim, not as our finding.
No pay-for-verdict
Companies cannot pay for a review, a placement in a list or a particular conclusion. We do not accept payment in exchange for coverage. If we later earn affiliate commissions, the commission rate will never decide which product we recommend, and pages with affiliate links will be clearly marked. See the affiliate disclosure.
Why there are no star ratings
We do not publish star ratings, scores or awards. A single number hides the trade-offs that matter, and we would rather show our reasoning against stated criteria.
AI companion coverage
Some AI companion products are intended for adults. Our coverage of them is non-graphic and focuses on cost, privacy, safety and value. Articles about adult-only products carry an 18+ notice.
Corrections and updates
If we get something wrong, we fix it and show an updated date on the article. Significant corrections are noted in the article itself. To report an error, email hello@aitriedhq.com with the page address and what needs checking.