Editorial policy

This policy describes how content on the site is selected, written, verified and corrected. It exists so that readers can judge not only what we say, but how we say it. The methodology page covers the concrete process of evaluating tools; this page explains the editorial thinking behind it.

1. Selection criteria

We cover AI tools that meet all of these conditions at once:

• They genuinely exist and can be used. We do not write about announced products with no release date, or abandoned projects.
• There is verifiable official information. If a provider publishes no pricing, terms or documentation, there is nothing to check against and we do not publish.
• They matter to someone in Spain. We focus on the local context: prices in euros, VAT, invoicing, GDPR and EU AI Act compliance.
• They add something for someone who already knows. We do not write tutorials on what a chatbot is. If a topic has been covered a hundred times with the same definition, we do not add another article.

2. How we write

Articles are produced with the help of artificial intelligence tools as a drafting and organising aid. It is worth stating that plainly rather than hiding it. Using AI as scaffolding does not absolve anyone: whoever writes is responsible for every published sentence.

In practice that means:

• No number is published unchecked. Prices, quotas, usage limits, dates and market figures are verified against the provider's official documentation, with the date checked, and the article links to that source.
• No claim is accepted because it sounds plausible. If it cannot be checked, it is not written, or it is written explicitly as a hypothesis.
• Data is separated from opinion. Judgement calls carry their reasoning so readers can argue with them. Verifiable statements carry their source.
• A person finishes and approves the copy. The name at the end of each article is the person accountable for it, and can defend its content.

Google's policy on AI-generated content does not ban the use of AI: it bans content generated mainly to attract clicks, republished without added value, or mass-produced to manipulate rankings. That is exactly what this site does not do, and it is also what makes the site unviable: AI model prices do not change because of an article, they change because the provider decides.

3. Authorship

The site is written and maintained by a single person, who is credited as the author on every article and is the site owner. There are no pseudonyms, no invented profiles and no signatures from experts who do not exist. The authors page reflects that without embellishment.

If an article by someone else is ever added, that person will have their own page with their full name and will review what is published under their byline before it goes out. Nothing is published under a signature that person has not approved.

4. Independence and conflicts of interest

The site is funded by advertising. We do not accept payment or collaboration in exchange for publishing, or for not publishing, a review. No AI provider pays to appear on this site, and none pays for a favourable rating.

This has an uncomfortable consequence and we prefer it: when a tool seems bad to us, we say so even if there are advertisers from the same category on the page. If paid content ever appeared, it would be labelled as such visibly, both in the article and on this page, and that is the only acceptable way to do it.

This site uses no affiliate links. When we link to a provider's documentation or pricing page, it is because it is the primary source you need to check, not because we get paid for the click.

5. Sources and verification

Every article declares its sources and the date they were consulted. We always prefer primary sources —official documentation, pricing pages, the provider's terms— to second-hand coverage. Where only secondary sources exist, we say so, because a chain of five articles citing each other is not verification: it is rumour.

When we cannot check something, the article says so. It is better to publish "we have not been able to verify that this feature is available on the free plan" than a confident claim that turns out to be false two months later.

6. Updates and expiry

Every article shows its publication date and, when reviewed, its last update date. We prioritise reviewing content that expires fastest: prices, usage limits, integrations and changes in provider policies.

When an article becomes outdated and we cannot update it with confidence, we flag it as such rather than leaving it published as if still current. Outdated information without a warning is worse than publishing nothing, because it looks true.

7. Corrections

Corrections are logged on the corrections page, with the date, what changed and why. If an error could have led someone to a wrong decision, the correction is extended to any article that depended on the wrong figure, not just the one containing the mistake.

If you find an error, write to alvarocimadevillawebs@gmail.com. Corrections reported by a reader carry the same weight as the ones we find ourselves.

8. Changes to this policy

This policy may be updated. The last review date appears at the top and changes apply to the version published at that moment.