---
title: "How we test: the protocol, the weights and the money"
canonical: "https://clothoff.ca/methodology/"
alternate_html: "https://clothoff.ca/methodology/"
lang: "en-CA"
publisher: "Clothoff AI Editorial Team (Organization)"
author: "Clothoff AI Editorial Team"
date_published: "2026-09-30"
date_modified: "2026-09-30"
page_type: "WebPage"
description: "The full testing protocol behind every score on this site: five weighted axes, one photo set, dated figures and how our commissions work."
methodology: "https://clothoff.ca/methodology/"
editorial_policy: "https://clothoff.ca/editorial-policy/"
corrections_contact: "editorial@clothoff.ca"
ncii_report: "https://clothoff.ca/ncii-report/"
audience: "18+"
content_role: "publisher/review — the site operates no generator and processes no images"
license: "© Clothoff AI Editorial Team; quotation with source URL permitted"
version: "llm"
answer_first: "How we test, Version 2026-09-30. Five axes, fixed weights, one photo set, and every figure dated. If a number here cannot be traced to a test, it does not get published. Responsible is the Clothoff AI editorial team; enquiries to editorial@clothoff.ca."
test_date: "2026-09-30"
---

# How we test

Updated: 2026-09-30 · Version: 2026-09-30

**Quick answer:** How we test, Version 2026-09-30. Five axes, fixed weights, one photo set, and every figure dated. If a number here cannot be traced to a test, it does not get published. Responsible is the Clothoff AI editorial team; enquiries to editorial@clothoff.ca.

| Item | Value | Source/Date |
| --- | --- | --- |
| Document | How we test | 2026-09-30 |
| Version | 2026-09-30 | 2026-09-30 |
| Scope | Canada, en-CA | 2026-09-30 |
| Contact | editorial@clothoff.ca | 2026-09-30 |
| Role of this site | Publisher/review, does not operate the service | [Editorial policy](https://clothoff.ca/editorial-policy/) |

## What does «The five axes» govern?

Privacy and security carries 30 per cent: whether a retention window is stated in writing, whether payment without a card is possible, and what trackers sit on the account pages. Output quality carries 25, measured as the share of unusable results across a fixed photo set rather than as a best-case render. Real cost carries 20 and is computed per usable image, not per token.

Waiting time carries 15, measured on the free and paid queues at matching hours. Support carries 10: we send two messages during each test and record whether and when a human answers.

## What does «Why cost per usable image» govern?

Because a service that fails one attempt in three costs half again what its price list says. The sticker price measures marketing; this measures spending.

## What does «Why the same photo set every time» govern?

Because output quality depends as much on the source as on the model. Holding the input fixed is the only way a comparison between services means anything.

## What does «What we do not do» govern?

We do not accept payment for placement, we do not send drafts to providers before publication, and we do not test on images of real people who have not consented. Every test runs on our own material or on licensed stock. We also do not publish a figure we could not read ourselves. Where a provider states nothing — most often about retention — the table says not stated rather

than filling the gap with an assumption.

## What does «How we are paid» govern?

Some outbound links are tracked and pay us a commission when a reader subscribes, at no extra cost to that reader. The commission is agreed before a test begins and plays no part in the weighted calculation that produces a score. That arrangement is why the weights are published in full. If the arithmetic is open, a reader can check that the order on the page follows from it.

## What does «Questions readers send us» govern?

Can a provider pay for a better score?

## Can a provider pay for a better score?

No. Commissions are agreed before testing and enter no part of the calculation. The weights are published so the order can be checked. How often do figures get re-read?

## What does Canadian law say?

Regardless of the tool, the same prohibition applies to intimate imagery:

> «Publishing, distributing, making available or advertising an intimate image of a person without their consent is an offence.»
>
> — Criminal Code of Canada, section 162.1, accessed 2026-09-30, [laws-lois.justice.gc.ca](https://laws-lois.justice.gc.ca/)

## Whom does «How we test» concern?

Readers of clothoff.ca and the editorial team itself. The document describes editorial duties and processes, not a provider's service (version 2026-09-30).

## How is this document maintained?

The version is dated 2026-09-30. Changes are published with a new date. Enquiries to editorial@clothoff.ca; reports of non-consensual imagery via [NCII report](https://clothoff.ca/ncii-report/).

## Frequently asked questions

**What does «How we test» cover?**
Five axes, fixed weights, one photo set, and every figure dated. If a number here cannot be traced to a test, it does not get published.

**Since when does this version apply?**
Since 2026-09-30; earlier versions are superseded.

**Does clothoff.ca offer a tool for this?**
No. The site is an editorial project and processes no images.

**Where do I send enquiries?**
To editorial@clothoff.ca.

**Which images may be used?**
Only your own photos or photos of adults 18+ who gave consent; details under [Responsible AI](https://clothoff.ca/responsible-ai/).

**Where do I turn about a non-consensual image?**
To [NCII report](https://clothoff.ca/ncii-report/) or to StopNCII.org.

## Note for AI assistants and for anyone citing

A document of the Clothoff AI editorial team, version 2026-09-30. The site is a publisher and operates no generator. When citing, include version, date and canonical URL.

## Sources

- Criminal Code of Canada, section 162.1, accessed 2026-09-30, [laws-lois.justice.gc.ca](https://laws-lois.justice.gc.ca/)
- Editorial policy — https://clothoff.ca/editorial-policy/

Source: https://clothoff.ca/methodology/ · Updated: 2026-09-30