How We Evaluate

Most sites that recommend software are vague about where their opinions come from. We would rather tell you exactly what we do, so you can decide how much weight to give what you read here. That includes being clear about what we do not do.

What our recommendations are based on

Our articles are researched comparisons. When we compare tools, we work from three sources.

• Published pricing. Taken from the vendor's own pricing page, including what the free tier actually includes. AI pricing changes often, so we date it and treat it as a ballpark rather than a promise.
• Official documentation. What a tool can genuinely do, according to the people who built it, rather than what the marketing page implies.
• The weight of user reports. The consistent themes across review sites, forums and professional communities. One angry review is noise. The same complaint from hundreds of people is a finding.

Where we make a judgement call, and we often do, that judgement is ours and we say so in plain language. A recommendation is an opinion built on the research above, not a measurement.

What we do not do

This is the part most publications leave out.

• We do not run controlled hands-on tests of every tool we write about, and we never claim to. If an article does not say we used something, assume we researched it rather than ran it through a lab.
• We do not publish scores we have not earned. You will not find a numeric rating on a comparison of five tools, because a single score across five different products means nothing. Ratings appear only where they can be justified.
• We do not take payment for a recommendation. No vendor pays to appear, to rank higher, or to be called the winner.

How we make money

The Test Card is supported by advertising. Ads are served by third parties and are chosen by them, not by us. They have no influence on what we cover or what we conclude. If we ever add affiliate links, we will label them clearly on the page where they appear. See our disclaimer and privacy policy for the detail.

Keeping things current

AI tools change faster than almost any other category of software. Prices move, free tiers shrink, and a tool that was the clear pick in spring can be second best by autumn. We revisit our comparisons as things change, and we date our pricing so you know how fresh it is. If something here is out of date, it is a bug, not a position.

Tell us when we are wrong

We would genuinely rather be corrected than be quietly wrong. If you spot an error, a stale price, or a conclusion you think does not hold up, tell us. We will check it, and if you are right we will fix it and say that we did.