Why this page exists
Most AI tool directories are built from vendor marketing pages. We think that is close to useless when you are about to spend money, so this page states plainly what we do, what we check, and — importantly — what we have not done for any given page. If a claim on a tool page is not backed by the method below, it does not belong on the page.
Who runs these tests
Every review on this Site is written and signed by Kelen, the site's founder and editor. There is no anonymous content team and no outsourced writing — the person named on a tool page is the person who checked its pricing, used the tool, and wrote the assessment. Where a claim has not been verified by hand, the page says so rather than implying a test that did not happen.
1. Pricing is verified against the vendor's own page
Every price, free-tier limit, seat count and plan name we publish is read from the vendor's official pricing page or documentation, not from a press release summary or another directory. Each tool page records the date that check was performed. Because AI pricing changes constantly — plan names get renamed, free tiers shrink, credits get repriced — a date is the only honest way to present a price.
2. Features are checked against documentation, not screenshots
Feature lists are assembled from official documentation, changelogs and release notes. We distinguish between what a tool claims and what we could confirm. Where we have only the vendor's word for a capability, the page says so rather than presenting it as established fact.
3. Hands-on use
For tools with a usable free tier or a no-card trial, we sign up and work through the same set of everyday tasks we would ask of a paid tool: produce something real, push it into a second step, and see where it breaks. That is how most of the limitations on our pages were found — they are the things that go wrong at the edges, not the things listed in a feature matrix.
We are straight about the limits of this: we do not run a lab, we do not benchmark models on standardised datasets, and we do not buy every enterprise plan. Where our judgement comes from documentation and a free-tier run rather than an extended paid trial, the page reflects that. We would rather tell you the basis of a claim than sound more authoritative than we are.
4. What we assess
- Capability — does it actually do the job it advertises, in normal daily use rather than a demo?
- Pricing clarity — is it obvious what you get for free and what triggers a charge?
- Free-tier honesty — is the free plan usable for real work, or a demo designed to expire?
- Friction — how many steps, sign-ups, credit cards and settings stand between you and useful output?
- Limitations — where it fails, what it refuses to do, what breaks at scale, and what the documentation leaves ambiguous.
- Alternatives — what else in the same category does the same job better for a specific kind of user, and at what price.
5. How we present results
We do not use a single numeric rating, because a single number hides the trade-off that actually matters when you are choosing. Instead each tool page states who the tool is for, who should skip it, what it does well, and where it falls short — followed by a direct comparison with the closest alternatives in our directory.
Category pages use the same standard across every tool in that category, so the comparison is like-for-like rather than a list of unconnected verdicts.
6. Re-checks and corrections
Price checks are re-run on a rolling schedule and at least quarterly, and each page shows the date it was last reviewed. Products that shut down, get acquired, or change their pricing model are updated or removed. If you spot something out of date or wrong, tell us through the contact page and we will correct it — substantive corrections are dated on the page itself.
Read our editorial policy for how we decide what to publish, and our advertising and affiliate disclosure for how the Site is funded.