Skip to content

How it works

  1. You write rules A policy is a short list of rules in plain English or Arabic, one idea per rule.
  2. The SDK captures ads Inside your Android or iOS app, the SDK finds ad views and captures each fully visible ad.
  3. AD80 receives them The SDK uploads the ad image and its metadata with your upload key.
  4. Each ad is scored AD80 scores the creative against every rule of the app's policy.
  5. You see and act Violations appear in the dashboard, the API, e-mail and webhooks.

Each app has one active policy: a list of rules written in natural language, in English or Arabic. Every save creates a new policy version. You can try a policy on screenshots in the Policy Playground before any app is connected.

2. The SDK captures the ads your users see

Section titled “2. The SDK captures the ads your users see”

The AD80 SDK starts when your app starts (on Android by itself, on iOS when your app calls Ad80.start). Once your app grants consent, it looks for ad views on screen (AdMob banners, Ad Manager banners and native ad views) and captures an ad only when it is fully visible and not covered. It copies only the pixels of the ad view, never the whole screen.

Each distinct ad image is uploaded once and later sightings are sent as small metadata records. See Data and privacy for exactly what is sent.

Uploads are authenticated with the app’s upload key (ingest key). Each new observation counts as one monitored impression for your plan. When the monthly allowance is reached, uploads are still accepted but the new sightings are stored as Not classified: monthly allowance reached; they are never shown as clean.

4. Each creative is scored against your rules

Section titled “4. Each creative is scored against your rules”

An AI model looks at the creative and gives each rule a score from 0 to 1, with an explanation for the rules it breaks. A creative that breaks at least one rule is a violation. When you save a new policy version on a plan with the SDK, the ads seen in the last 7 days are scored again against the new rules.

  • The Violations inbox lists ads that broke at least one rule. You Confirm violation or mark it a False positive.
  • Notifications send an e-mail or a signed webhook when a creative is classified or violates a rule.
  • On Enterprise, each violating creative opens an incident that you track to resolution.
  • Everything is also available through the REST API.