Our method for testing and comparing tools

Why we publish our method

A comparison written by one of the brands being compared starts out biased. We can’t remove that, but we can make it visible: we publish how we test so you can rerun the test and reach your own conclusion.

The method applies to three kinds of page:

• head-to-head comparisons between Aithor.io and another tool;
• lists of tools for a specific task, such as writing a dissertation;
• the short tables on our tool pages, which compare Aithor.io with a generic AI chat without naming any brand.

How we run each test

1. We fix the task and the criteria before testing. For example: “write the introduction to an undergraduate dissertation on sleep and academic performance, with three sources in APA 7”. The criteria don’t change once we have seen the results.
2. Every tool gets the same brief. We write it ourselves, in the language of the page, and it is never a real student’s work.
3. We record the plan, the version and the date. We test the free plan and, if the comparison covers the paid plan, that one too. If a tool lets you choose a model or a mode, we note which.
4. We keep everything it produces. Screenshots and the full text of each output, so we can show them and check them later.
5. We check every reference. We look each one up in Crossref ^[1] and OpenAlex ^[2], and open the source. A reference that doesn’t exist, or doesn’t say what it is cited for, counts as an error.
6. We score against the criteria and write the verdict. The verdict says where each tool comes out ahead, even when that tool isn’t Aithor.io.
7. The page passes our editorial review before it goes live. Like every page on the site, it goes through the automated checks described in our editorial policy: product facts, sources, links and internal consistency. If it fails, it doesn’t go live.

The criteria

Each comparison states which criteria it uses and why. These are the ones that come up most for academic work:

What we publish from each test

• the test date and the plan we used for each tool;
• the full test brief, so you can rerun it;
• real screenshots of the output;
• at least one reference with a DOI that you can check yourself;
• a link to each tool’s official pricing page, with the date we checked it.

How often we rerun tests

AI tools change fast. We review every comparison every three months and take it down if its data goes stale and we can’t rerun the test. The date of the last check always sits under the table.

What we don’t do

• We don’t put Aithor.io first by default: the ranking comes from the criteria.
• We don’t copy ratings or stars from other sites.
• We don’t use other brands’ logos.
• We don’t publish accuracy figures for an AI detector without our own published measurement.
• We don’t name another brand in a tool page’s table unless every data point comes from a dated test, with the plan and a link to its source.

If you think we got something wrong

If you work on a tool we compare and something isn’t right, email hello@aithor.io with your source. We check it, correct the page if needed and note the change with its date, as our editorial policy explains. Published comparisons are listed under comparisons.