How we work

NodoAI
editorial methodology

How we select, analyse and score the AI tools we review. No shortcuts: verification against official sources and hands-on use whenever possible.

What everything we publish is built on

Real testing, not demos

We never base a review on marketing materials, guided demos or press releases alone. We check every feature and price against official sources and, whenever possible, test the tool on real tasks before publishing.

Objective criteria

The final score is calculated from six dimensions with defined weights (see below). It is not a subjective impression: it is the result of a rubric applied equally to every tool.

Commercial independence

Affiliate relationships do not influence the scores. A tool with an affiliate link can get a 5/10; one without can get a 9/10. The two systems are completely independent.

Regular updates

AI tools change fast. We revisit reviews when there are significant product updates, price changes or new features that alter the evaluation.

How we evaluate an AI tool

01
Initial selection

We identify the tool through mentions in specialised communities, relevant launches or reader requests. We check it has a large enough user base to justify a review.

02
Sign-up and setup

We create a real account. If the tool has a paid plan, the subscription is paid by NodoAI to guarantee full access to every feature we evaluate.

03
Hands-on evaluation

Whenever possible, we use the tool for real use cases (creating content, analysing documents, coding or automating flows depending on the category) and document results, errors and unexpected behaviour. What we cannot test first-hand, we verify against official sources.

04
Rubric evaluation

We apply the six evaluation dimensions (see below) with their weights to calculate the final score out of 10.

05
Market comparison

We place the tool within its category: what alternatives exist? When is this tool preferable to its competitors? Which type of user is it best suited to?

06
Publishing and follow-up

We publish the review with a visible last-updated date. We monitor the product and update the review when changes affect the score.

The 6 evaluation dimensions

30%
Output quality

Accuracy, consistency and real usefulness of the outputs in real-world use cases. It carries the most weight because it is what matters to the end user.

20%
Ease of use

Learning curve, interface clarity and the time needed to get useful results with no prior experience.

20%
Value for money

We assess what you get for what you pay compared to alternatives in the same price range. An expensive plan can still score well if the results justify it.

15%
Speed and reliability

Response times, service stability, error rate and availability across the testing period.

10%
Integrations and API

Ability to connect with the rest of your work stack. We especially value native integrations, a documented API and automation compatibility.

5%
Support and docs

Quality of the official documentation, support response time and activity of the user community.

How to read the final score (out of 10):

9.0 – 10 Absolute reference A tool that sets the standard for its category. Few or no competitors beat it.
8.0 – 8.9 Highly recommended Solid, mature and with good value for money. Fits most use cases.
7.0 – 7.9 A good option Works well but has relevant limits in price, support or features.
6.0 – 6.9 Acceptable with caveats Useful in very specific contexts. Not the best option for general use.
5.0 – 5.9 Very specific use Only justified if no alternative covers your particular use case.
< 5.0 Not recommended Serious problems in quality, price, privacy or usability that outweigh its strengths.

Scores are calculated by weighting the six dimensions with the percentages shown and rounding to one decimal. Our directory clusters between 7 and 9: an 8.5 already means an excellent tool for professional use.

What we don't do

We do not accept payment for reviews or for improving a score. Ever.

We do not publish sponsored content disguised as an independent review.

We do not invent statistics or benchmarks: any figure has a verifiable source or it does not go in.

We do not hide the downsides. When a tool disappoints us, we say so — that is why our reviews have a "cons" section and a "when NOT to use it".

We do not let affiliate links decide the ranking. If we earn a commission, we disclose it clearly; it changes nothing about the verdict.

See our affiliate policy →