AdOverlayProof

Who it's for

Indie mobile app developers, QA/press teams, and consumer-rights reporters who today cope by manually saving screen recordings, screenshots, and scattered device logs (or by asking users to send videos via Slack/email).

The problem

Time + legal/reputation risk: teams waste hours assembling credible evidence to contest platform-inserted ads, or fail to prove an insertion when requesting refunds/press coverage or filing platform complaints — leading to lost revenue, delayed app updates, or PR damage.

How to build it

A tiny web app + single-file CLI uploader that accepts MP4/HEVC screen recordings or PNG/JPEG screenshots; produces a downloadable ZIP: signed PDF report, 3s clipped MP4, per-frame annotated images, and a short verification JSON that a verifier can re-check against the file hashes. Optional lightweight JS snippet for embedding a 'report an injected-ad' button in bug trackers.

How it makes money

Who pays: indie app teams, QA consultancies, and tech journalists; Pricing: one-off evidence packet $29–$99 or $15/month for 10 reports; Why they pay: each packet reduces time-to-evidence from hours to minutes and is necessary to file credible platform disputes or press pieces; Why not free: a signed, tamper-evident PDF requires server-side signing (HSM) and a limited quota prevents abuse; free screenshot upload wouldn't be trusted in disputes.

Why it doesn't exist yet

Platforms (Apple/Google) don't publish a simple deterministic signature for their inserted overlays; incumbents (big testing suites) avoid building legal-grade, signed evidence tools because it requires device-level metadata, robust tamper-evidence, and a small legal/audit workflow — an indie can stitch together heuristics, OS-accessible metadata, and a simple HSM-signing flow to produce defensible packets quickly.

First users

The first 10 users will be mobile indie devs and Apple-review-focused shops who are currently seeing new persistent iOS ads and need a reproducible complaint packet (one-offs worth $29–$99) — plus a few tech journalists who will use it to publish reproducible reports that attract traffic and drive more signups.

Build size

2 people x 8 weeks: includes web UI, a small Go/Rust CLI uploader, heuristics for overlay detection (pixel/template matching), device/OS metadata extraction from recordings/screenshots, HMAC/HSM signing and PDF generator, and minimal payments/checkout. Excluded: on-device kernel modules, native iOS app for background capture, or automated device farm integration.

Biggest risk

If Apple/Google publish an official reporting API or start embedding cryptographic signed insertion-metadata in screenshots/recordings, this indie service becomes redundant; also large observability vendors could add signed-evidence features quickly.

Conditions for a hit (all 3 required)

• Signed PDF evidence packet: for any uploaded MP4/PNG the system emits a PDF signed by the service's ephemeral key containing device model (from video metadata), OS build string, upload timestamp, SHA256 hashes of artifacts, and a one-paragraph overlay verdict (detected overlay present/absent).
• Annotated visual artifacts: produces a 3-second clipped MP4 and per-frame PNGs with overlay bounding-box(es) and pixel-diff heatmap images showing stable overlay pixels across frames (verifiable by re-running the same clip).
• Verification JSON endpoint: for each packet the service publishes a JSON verifier URL that returns file hashes, detection parameters (template IDs + thresholds), and a signed statement so a third party can confirm the packet matches the uploaded files within 10 seconds.

How it's judged (in 6 months)

Product Hunt daily top 5 or GitHub 1,000 stars (whichever is first)(judgment date 2027-03-26)

AI self-confidence 55/100 — self-reported likelihood of meeting the criterion, not a business success rate