Editorial Policy
The rules the Clothoff AI Editorial Team follows for independence, sourcing, corrections, retests, and conflicts of interest. Updated: September 3, 2026.
Summary
- Reviews are written by the editorial team as an organization; no vendor pays for a review, a rank, or a placement, and no vendor sees a draft.
- Facts come only from official service pages fetched by the team with an access date, and from primary legal texts; no user reviews are published or quoted.
- Drafts may use AI assistance; every fact, price, measurement, and score is verified by an editor before publication.
- Corrections: editorial@clothoff.ai · acknowledgment within 5 business days · confirmed errors fixed within 10 business days with a dated note.
- Every score is retested at least every 90 days; each review shows its Tested date.
- Affiliate commissions from 6 partner tools fund the site and never change a score — see Affiliate disclosure.
- Only your own photos or photos of consenting adults (18+) may be used with any reviewed tool — see Responsible AI.
Policy at a Glance
Seven rules govern every page on this site. Each is expanded in the numbered sections below, with the deadlines the team commits to.
| Rule | Standard | Basis |
|---|---|---|
| Independence | No paid reviews, no paid placement, no vendor pre-review; partner status disclosed and excluded from scoring | Affiliate disclosure |
| Sourcing | Facts only from official service pages fetched by the team, each with an access date; primary law texts for legal claims | 16 CFR Part 465 (Oct 21, 2024) |
| User reviews | None published or quoted; the site relies on its own dated Editorial Test Notes | 16 CFR §465.2, §465.4 |
| AI assistance | Drafts may use AI assistance; every fact, measurement, and score is editor-verified before publication | Section 3 below |
| Corrections | editorial@clothoff.ai · acknowledgment within 5 business days · fix within 10 business days · dated note on the page | Section 4 below |
| Retest cadence | Every score retested at least every 90 days or on a trigger event; Tested date shown on every review | How we rate |
| Consent | Test imagery only from consenting adult editors or synthetic sources; consent note on every ranking | Responsible AI |
1. How Do We Stay Independent?
The editorial team decides which tools to test, how to test them, and what to publish. Vendors cannot buy a review, a rank, a badge, or a placement, and they do not see drafts. Requests to change a score are answered with the test note that produced it; a score changes only after a new test run, never after a negotiation.
Six of the 15 reviewed tools run partner programs that pay a commission after a reader subscribes. That relationship is disclosed above the first sponsored link, at each partner button, and in the footer of every page that carries one. Partner status is not an input to the scoring method, and the ranking on Undress AI rankings is ordered by editor score alone.
Where a vendor provides test access (a trial account or credits), the review's “How we tested” section says so. Access is accepted only when a paid account is unavailable to the public, and it is never accepted in exchange for coverage.
2. Where Do Our Facts Come From?
Every factual claim about a tool comes from a page the team fetched itself: the official pricing page, terms of service, privacy policy, help center, or the product interface during a test run. Each is recorded with its URL and access date in the review's Editorial Test Note and in the Sources list. Marketing claims, app-store descriptions, and forum threads are not treated as facts.
Legal statements cite primary texts: the United States Code, the Code of Federal Regulations, official regulator guidance, and EUR-Lex for European law. Where the team cannot verify a claim, the review says “not documented” rather than guessing.
The site publishes no user reviews, star ratings, or testimonials, and quotes none from third-party sites. The FTC Trade Regulation Rule on the Use of Consumer Reviews and Testimonials (16 CFR Part 465), effective October 21, 2024, prohibits fake, purchased, and insider reviews (Federal Register, Aug 22, 2024). The team cannot verify the origin of anonymous reviews, so it publishes only its own dated test notes.
3. Do We Use AI to Write Reviews?
Drafts may use AI assistance for outlining, language editing, and translation of source documents. AI is not a source: every fact, price, timing measurement, and score in a published review is checked by an editor against the fetched source or the team's own test log before publication. A claim that cannot be traced to a source is removed, not softened.
Sample imagery shown on review pages is AI-generated output from the reviewed tool, produced from the team's consented control images, and is labeled as such in the caption and in the alt text. The labeling rules follow the AI content policy and mirror Article 50(4) of the EU AI Act, which applies from August 2, 2026:
“Deployers of an AI system that generates or manipulates image, audio or video content constituting a deep fake, shall disclose that the content has been artificially generated or manipulated.”
— Regulation (EU) 2024/1689 (AI Act), Article 50(4), accessed September 3, 2026, eur-lex.europa.eu
4. How Do We Correct Mistakes?
Anyone can request a correction by emailing editorial@clothoff.ai with the page URL, the statement in question, and a source. The process has fixed deadlines:
- Acknowledgment within 5 business days, with a case reference.
- Verification against the original source and, where needed, a fresh fetch or test run.
- A confirmed factual error is fixed within 10 business days; the page receives a dated correction note and an updated dateModified.
- A disputed claim that cannot be settled within 10 business days stays online with a visible “under review” note until the check is complete.
- Score changes that result from a correction are published with a new Tested date, never silently.
Requests that concern a person's image rather than a fact go to ncii@clothoff.ai or the Report NCII form and are handled first, with a 48-hour response target. Copyright claims follow the DMCA procedure. Corrections handled each quarter are counted in the Transparency report.
5. How Often Do We Retest?
Every editor score is retested at least every 90 days. Each review displays a Tested date next to the score; when a score is older than 90 days, the page marks it as pending retest until the new run is published. The test protocol is identical for the first test and every retest, so scores remain comparable over time.
A retest is brought forward when the tool changes its pricing, privacy terms, retention window, or removal route; when the output engine changes visibly; when the service is unreachable for more than 24 hours during monitoring; or when a reader reports a discrepancy that the team confirms. Category pages such as All reviews inherit the latest scores automatically.
6. What Are Our Conflict-of-Interest Rules?
Nobody on the editorial team holds equity, employment, or a consulting role with any reviewed vendor, and the team accepts no gifts, paid travel, or payments from vendors. Test access provided by a vendor is disclosed in the review. Commissions from partner programs are paid by the vendor after a reader subscribes; they are the project's funding source and are stated on every page that carries a sponsored link.
Commissions never change scores. The scoring method fixes the weights (privacy 30, quality 25, price 20, speed 15, support 10) before a tool is tested, and the same protocol applies to the 6 partner tools and the 9 non-partner tools. A partner tool can and does rank below a non-partner tool when its measurements are weaker. The disclosure standard is the FTC rule on material connections, 16 CFR §255.5 (law.cornell.edu); the practical rules are in the Affiliate disclosure.
7. What Consent Standards Apply to Test Imagery?
Only use photos of yourself or of adults (18+) who have given explicit consent. Creating or sharing intimate images of a real person without consent is a federal crime under the TAKE IT DOWN Act. See our Responsible AI policy.
The editorial team applies that rule to itself. Test runs use a fixed set of control images: photos of consenting adult editors and synthetic images with no real subject. No photo of a third party, public figure, or minor is ever uploaded to a reviewed tool, and control images are deleted from the tool's account at the end of each run where the interface allows it.
Any tool that produces imagery of minors is delisted and reported to the NCMEC CyberTipline (report.cybertip.org). The same happens to a tool that offers no removal route or that ignores a consent report. The TAKE IT DOWN Act (Public Law 119-12) sets a 48-hour removal duty for covered platforms; this site applies the same 48-hour response target to reports received through Report NCII.
Contacts
- Corrections and editorial questions: editorial@clothoff.ai (acknowledgment within 5 business days)
- Legal notices: legal@clothoff.ai · Copyright: dmca@clothoff.ai
- Non-consensual imagery: ncii@clothoff.ai · Report NCII (48-hour response target)
- Privacy requests: privacy@clothoff.ai · Data request
All role addresses and response targets are listed on the Contact page. The project has no public office; notices are accepted by email.
Related Policies
- How we rate — weights, sub-criteria, and the test protocol
- Affiliate disclosure — partner programs and sponsored-link labeling
- AI content policy — labeling of AI-generated sample imagery
- Responsible AI — consent rules and safety commitments
- About us — who runs the project and how it is funded
- Transparency report — corrections, takedowns, and delistings by quarter
Frequently Asked Questions
Do vendors pay for reviews or placement?
No. No vendor pays for a review, a rank, or a placement, and no vendor sees a draft before publication. Six reviewed tools have partner programs that pay a commission after a reader subscribes; that status is disclosed on the page and never enters the score, as the Affiliate disclosure explains.
Are the reviews written by AI?
Drafting may involve AI assistance for structure and language, but every fact, price, measurement, and score is verified by an editor against the source before publication. No claim is published because a model produced it. Sample imagery from reviewed tools is AI-generated and labeled under the AI content policy.
Why does the site publish no user reviews?
Because the site cannot verify who wrote them or whether they were bought. Since October 21, 2024 the FTC Consumer Reviews Rule (16 CFR Part 465) bans fake and purchased reviews; the simplest way to comply is to publish only the team's own dated test notes and to name every source.
How fast are corrections handled?
A correction request to editorial@clothoff.ai is acknowledged within 5 business days. A confirmed factual error is fixed within 10 business days, and the page receives a dated correction note plus a new dateModified. Disputed claims stay online with a visible note until the check is complete.
When does a review get retested?
Every score is retested at least every 90 days, and earlier if the tool changes its pricing, privacy terms, removal route, or output engine, or if a reader reports a discrepancy. Each review shows a Tested date; a score older than 90 days is marked as pending retest until the new run is published.
What happens if a tool violates the consent rules?
The tool is delisted, not scored. Delisting applies when a tool produces imagery of minors, offers no working removal route, ignores a consent report, or markets itself for use on real people without consent. The delisting and its reason are recorded in the Transparency report.