How does photo-verified cleaning work?

What the cleaner does

The job link on the cleaner's phone lists the property's photo areas. Tapping an area opens the camera inside the page (no app), the cleaner takes the photo, sees it next to the property's reference photo for that area if the host uploaded one, and either uses it or retakes. If the phone blocks the camera — common on iOS when permission was once denied — the page explains the fix and offers the photo gallery as a fallback; a blank capture is detected and rejected rather than uploaded.

Photos are required before the job can be marked finished when the host has switched "require photos" on for that cleaner. A host can also require the checklist and require photo approval before the job counts as done.

What the host sees

• A notification that photos are in, on whichever channels the host chose.
• A photo review page per job with each area, its status (pending, approved, rejected), and a comment thread per photo. Co-hosts and property managers can review too; comments record who wrote them.
• Approve all in one tap, or reject a specific area with a note. Rejection reaches the cleaner while they are still at the property.

Why per-area rather than "a few photos"

A free-form photo requirement produces three pictures of the same clean sofa. Naming the areas makes the omission visible: no bathroom photo means the bathroom is unverified. It also makes the photos comparable week to week, which is what a reference photo is for.

Numbers

916 photos across jobs since 2026-08-11, when in-page camera capture replaced the native capture flow; 152 checklist reports submitted since February. Source: feed integrity report.

Technical details

The photo notification is photos_submitted.