Methodology

The AI sector changes so fast that the main risk of publishing about it is not being wrong but ageing badly: an article that was correct six months ago may no longer be. This page explains what we check and by what criterion we decide, so you can weigh every figure you read.

1. The four checks

1. Documentary verification. This is the basis of everything. A price, quota, feature or condition we state has to appear in the provider's official documentation, and the article links to it with the date we checked. With no official source, the figure is not published.
2. Practical checking. When a tool is paid or has a free plan, it is used long enough to see whether it behaves as advertised. Reading the marketing page is not enough.
3. Comparison with the alternative. A price only means something in comparison. Before recommending something, we check what the nearest alternative offers at the same price, because the buying decision is almost never between one option and none.
4. Spanish context. A price in dollars is not the price paid in Spain. For each figure we check VAT, the invoicing method if you are self-employed or a company, whether a card purchase requires a contract, and whether the provider complies with the GDPR and the EU AI Act. This is the part of the analysis English-language comparisons skip.

2. Criteria for deciding what to publish

A topic enters the site if it meets all four conditions in the editorial policy: the tool exists and can be used, there is verifiable official information, it matters to someone in Spain, and it adds something for someone who already knows the basics. If any fails, we do not publish, even if there is search demand.

3. How pricing is maintained

Model prices are stored in a single file in the code and every page reads from there. That decision has a practical consequence: correcting a price means changing a number in one place rather than editing twenty articles, which is why we can offer a reliable verification date on the pricing page.

The token cost calculator applies those same prices to the values you enter, so you can reproduce the calculation. It is a way to check whether the figure we gave you is right.

4. Independence

Ratings are independent. We do not accept payment for reviews, we do not let a provider review text before publication, and we use no affiliate links. If an AI tool seems bad to us, we write it, even if there are advertisers from that category on the same page. The full explanation is in the editorial policy.

5. Limitations worth knowing

It would be dishonest to present these pages as a definitive reference without more. There are limitations a reader should keep in mind:

• The analysis is done from a single point of view: that of a user and a business in Spain. We do not speak for other sectors or other countries.
• Pricing information is accurate on the date checked and is inherently perishable. Any purchasing decision must be verified on the provider's own page.
• We do not assess the security of tools in depth as an auditor would. Where an article touches security, it relies on the provider's documentation and does not substitute for an audit.
• We do not estimate usage, quota or performance figures the provider does not publish: they are not written.

If you find a flaw in the method or a figure that does not hold up, write to alvarocimadevillawebs@gmail.com. Corrections are published on the corrections page.