Published scorecard · checked 2026-09
How we review
Published · Updated · by the AISexGenerator.org editorial desk
Every score on this site comes from the same eight criteria, weighted the same way, applied to the same kind of evidence. This page is the whole method. If a number on a review cannot be traced back to something written here, it is a mistake and we want to hear about it.
Our evidence: public pages, dated
We compare services using what anyone can check without paying: pricing pages, feature pages, public galleries and demos, help-centre articles, terms of service and content policies. We record each fact with the date we read it — the current round was checked in September 2026. We do not create accounts, buy plans or enter personal details to gather evidence, so our reviews say “compared” and “reviewed”, never “tested”.
That choice has a cost we would rather state than hide: we cannot measure how quickly a paid plan renders, how many attempts a difficult pose takes, or how strict a filter is in practice. Where a fact lives behind a sign-up, the review says it was not verifiable and when we looked.
The eight criteria and their weights
Each criterion is scored from 1 to 10. The overall score is the weighted average, rounded to one decimal. A 10 is not “perfect”; it means the service does everything that criterion asks for and documents it publicly.
Generation range 20%
Which output types the service publicly lists: still images, animated or video clips, prompt-driven generation, scene or pose presets.
A 10 looks like: Images and video from both prompts and presets, all documented on public pages.
A 7 looks like: Images only, or video mentioned without any published detail about how it is made.
Character control 15%
How precisely you can shape a character: face, body, hair, ethnicity, outfit, style (realistic or anime), and whether a character is saved for reuse.
A 10 looks like: A step-by-step builder with saved characters that stay consistent between generations.
A 7 looks like: A text box and a few presets, with no stated way to keep a character consistent.
Explicit scope and clarity 10%
Whether the service states plainly what adult content it permits, from nudity to explicit scenes, and what it blocks.
A 10 looks like: Explicit output clearly permitted, with the blocked categories listed in the terms.
A 7 looks like: Adult content implied by marketing but not described in the rules.
Pricing transparency 15%
Whether prices, billing periods, currencies and credit costs per image or clip are published before sign-up.
A 10 looks like: Every tier and the per-image or per-clip credit cost visible on a public page.
A 7 looks like: Prices only shown after registration, or credits with no published cost per output.
Value for money 10%
The published monthly price against what it includes, compared with the other services in the same review round.
A 10 looks like: An entry plan at or below the group median that still includes image generation.
A 7 looks like: An entry plan well above the group median, or one that leaves out the core generator.
Trial access 10%
What a visitor can do before paying: free messages, free images, a trial, or nothing.
A 10 looks like: A stated free allowance that includes at least one generated image.
A 7 looks like: Sign-up possible, but every generation feature locked until you pay.
Privacy and billing discretion 10%
What the privacy policy says about data use and deletion, and whether billing descriptors are described as discreet.
A 10 looks like: Account and data deletion described, discreet billing stated.
A 7 looks like: A generic privacy policy with nothing specific about generated content or billing.
Safety rules 10%
A published 18+ requirement and bans on sexual content involving minors and on real people's likeness without consent.
A 10 looks like: All three rules stated in the terms or content policy, plus a way to report abuse.
A 7 looks like: An age requirement only, with no published rule on real-person likeness.
What we could not verify
Four things sit outside this scorecard because public pages cannot prove them: the realism of paid output at full resolution, real rendering times, how often a filter blocks a legitimate adult request, and how billing actually appears on a statement. When a service makes claims about any of these, we quote the claim as the service's own and do not score it.
How scores stay consistent across the site
The number next to a service in the reviews hub, in its full review and in the comparison on the home page is one value stored once, so it cannot drift between pages. When a price or a rule changes, we update that value, the review text and the page date together.
How to challenge a score
Send the page, the criterion and a public source to contact@ai-sex-generator.org. If the source shows we were wrong, the score changes and the review's update date moves. More about the publication is on our about page.