CleanMy Moderation Policy

Effective date: to be set at publication. Published at: https://cleanmy.life/moderation.

This document is the source for the public Moderation Policy page. The apps/web/app/moderation/page.tsx Next.js route renders this content.

The technical reviewer-facing version is at ../03-attachments/moderation-policy-public.md.

How CleanMy moderates content

CleanMy lets adult and parent users publish their cleaning artifacts to a public community feed. This page describes how we keep that feed safe and what to do if you see something that shouldn't be there.

What CAN be published

CleanMy is for photos of physical spaces you've cleaned. Before-and-after shots of your kitchen counter, a kid's room, a desk, a closet. The artifact celebrates the transformation.

What CANNOT be published

We don't allow:

• photos with an actual live person in the shot. People DEPICTED as part of the room (framed photos, family photographs, posters, decor) are fine; a real person standing in the photo is not. CleanMy photos are of the space.
• nudity or sexually explicit imagery, whether a real person or shown in a poster, painting, magazine, or screen. Swimwear or ordinary clothing is fine; the line is explicit content. Our AI screens every photo before it can be saved.
• photos of identifying information (addresses, license plates, school IDs),
• pornographic, violent, or threatening imagery,
• content depicting minors that isn't a clean room they just tidied,
• harassment, hate speech, or threats,
• spam, ads, promotional content,
• AI-generated or stock photos submitted as "cleaning artifacts,"
• recycled photos used to fraudulently earn screen time.

Full terms at cleanmy.life/terms.

What people can post and write

The CleanMy public feed surfaces two kinds of user-generated content:

1. The cleaning artifact itself — before and after photos of a physical space, optionally with a short caption the publisher writes.
2. Public comments on other people's cleans — short, asynchronous replies (think the comment thread under a Reddit post, not a chat message). Comments are visible to anyone who can see the post.

Both surfaces are bound by the same content rules in the Terms of Service. Adults are the only people who can publish or comment in CleanMy V1; the kid surface cannot reach the feed at all and contains no posting or commenting UI.

How we filter before something is published

Every published photo passes through:

1. Client-side EXIF strip. Before upload, your photo is re-encoded as a fresh JPEG on your device, which removes EXIF metadata (including any GPS coordinates the camera would otherwise embed) as a side effect of the re-encode.
2. Pre-save AI safety classifier. Every photo that enters the AI-clean pipeline (both the before photo and the after photo) is examined by Google's Vertex AI for content that should not be processed: an actual live person physically in the shot, nudity or sexual content (whether a real person or shown in artwork, posters, magazines, statues, screens, or mirror reflections), any content that sexualizes a minor, graphic violence or injury, weapons (other than ordinary kitchen knives in a normal kitchen context), and drug paraphernalia. People merely DEPICTED as part of the room (framed photos, posters, decor) and clothed/swimwear figures are NOT flagged. The classifier runs in the same Vertex AI call that grades the clean, so rejection happens before any tasks or credit. On any rejection the photo is removed from your account immediately and held in a private, restricted store for our safety review (it is not published and is never left at a public link), no Clean credit is charged, and you are asked to retake (or use Quick Clean, which involves no photo at all). Because this runs on every clean, it protects photos whether or not they are ever published.
3. Kid-content always requires explicit parent review. A clean a child captures cannot become public on its own. It is delivered to the paired parent's review inbox, and the parent must take an explicit publish action (with a fresh Sign in with Apple confirmation tied to that specific photo) for it to appear on the public feed. The parent review step is the load-bearing consent gate, on top of the AI classifier above, which blocks nudity, sexual, or graphic content before it is ever saved.
4. Duplicate-photo fraud check. Server-side SHA-256 deduplication catches re-uploads of the same photo within a 48-hour window.
5. Report + Block + human review. Every post in the feed carries a Report Inappropriate Content button (described below). Reports queue for human review within 24 hours; substantiated reports result in deletion and account-level action.

Roadmap

The pre-save AI classifier above already blocks people and sexual content on every photo and is the load-bearing defense in V1. These additional layers are designed but not yet in production and add defense in depth:

• A dedicated face-detection pass as a separate stage from the classifier.
• Perceptual-hash matching against industry databases of known-illegal content (e.g. Microsoft PhotoDNA, Thorn Safer). The server integration point is wired; it activates once we complete vendor onboarding.
• A standalone vision-model NSFW classifier as an independent second pass on published content.

We will update this page when each ships.

How to report something we missed

Every post in the CleanMy feed has a Report inappropriate content button at the bottom. Tap it, pick a reason, hit submit.

What happens next:

• The post is immediately hidden from your view, regardless of whether anyone has reviewed it.
• The report queues for human review within 24 hours.
• If we find the report substantiated, we permanently delete the post and consider account-level action against the uploader (account suspension, account termination, in severe cases reporting to authorities).
• We do not tell the original poster who reported them.

Individual comments can be reported the same way, using the Report action on the comment itself. A reported comment is hidden from your view immediately and queued for the same human review.

How to block someone

Every public clean in CleanMy carries a Block this user action. Tap it and confirm. From that moment:

• All of that user's posts disappear from your view.
• The blocked user cannot see your posts.
• The block is permanent and cannot be undone from the app. If you change your mind, email support@cleanmy.life and we will reverse it on request.

Kid content (extra protection)

Children using CleanMy can produce cleaning artifacts but they cannot publish them. The only way a kid's clean reaches the public feed is:

• the kid sends their finished clean to the paired parent,
• the kid's clean appears in the parent's "Yours" tab on the parent's device, marked as pending review (a database rule forces every child-captured post into this pending state, so it can never publish on its own),
• the parent reviews the clean and taps "Add to my feed",
• the parent must do this on the parent's own Sign in with Apple verified session, consenting specifically to this clean being public,
• our server records the approving parent's verified identity and makes the parent the publisher of record, giving an auditable per-act consent record,
• the resulting public post is owned by the parent, not the kid, and is labeled with the parent's chosen public name (or a system-assigned user handle like user_a4b3c2d1, derived from a one-way hash of the parent's account ID, if the parent has not set one). The child's first name is never used as the public post label.

Our service-level commitment

During the public beta period, we review reports on a best-effort basis; we are notified by email the moment a report lands and review as quickly as practical. As CleanMy reaches scale we will commit to a 24-hour first-pass-decision SLA and publish performance against it.

If you have reported something and a few days have passed without visible action, email abuse@cleanmy.life.

Appeals

If your post was removed or your account was suspended, you can appeal by emailing abuse@cleanmy.life with your reasons. We review appeals within 7 days and respond with a final decision. We will overturn moderation decisions if we agree the original action was wrong.

Transparency

We will publish a transparency summary once CleanMy has meaningful public volume. The intention is to share what we track internally: total posts submitted, total reports filed, median time-to-first-decision, and total accounts actioned. No timeline commitment until we are past the public beta period.

Contact