Skip to content

Key concepts

These are the terms you will meet in the dashboard, the API and the rest of this documentation. The glossary lists them with their API names and their Arabic equivalents.

Your organization is your company’s account in AD80. People join it by invitation and have one of three roles: Owner, Admin or Viewer. See Roles and permissions.

An app is one Android or iOS application, identified by its app id: the Android application id (the package name) or the iOS bundle identifier, for example com.example.news. Apps hold a policy, upload keys and the creatives their users saw. Each organization also has a Policy Playground app that the dashboard creates for you and keeps out of the SDK lists.

A creative is one distinct ad image. AD80 identifies it by a content hash of its pixels (ad80px1: followed by 64 hex characters), so the same image seen many times is one creative with many sightings. A creative has a source: Reported by your SDK (sdk) or Uploaded by you (manual, from the Policy Playground).

A rule is one sentence that describes something you do not allow, for example “No ads showing alcohol”. Each rule has a short rule id (letters, digits, dot, dash or underscore, up to 64 characters) and an optional display name. A policy is the list of rules for one app.

Every save of a changed policy creates a new version (v1, v2, …). Saving identical rules creates no new version. Each score says which policy version it was made with, so you always know which rules were applied.

A classification is AD80 checking one creative against a policy. Its status is one of:

StatusMeaning
ClassifiedA verdict exists: every rule has a score.
Classification pendingWaiting to be scored.
Scoring failedThe check failed.
Not classified: monthly allowance reachedThe sighting was counted but not checked, because the plan’s monthly monitored-impression allowance was used up. It is never shown as clean.

The verdict of a classification gives each rule a score between 0 and 1 and says whether the rule is violated. It also carries a short description of what was seen and an explanation of why a rule was broken.

A creative whose latest verdict breaks at least one rule. Violations appear in the Violations inbox.

You review a violation: Confirm violation when AD80 is right, or False positive when it is not. The review status is Unreviewed, Confirmed or False positive. Your answers are kept as feedback and are used to check how well the scoring works.

An incident tracks one violating creative that your SDK reported, from the first sighting to resolution, with the status Open, Investigating, Resolved or Dismissed, notes and a resolution history.

An upload key is the secret the SDK uses to send ads to AD80. It starts with ad80_ik_, belongs to one app and can only upload. You put it in your app: the manifest on Android, Info.plist on iOS. The API calls it an ingest key.

An API key is the secret your scripts and servers use to call the REST API. It starts with ad80_ck_ and carries scopes that limit what it can do. Never put an API key inside an app.

Monitored impressions, manual classifications, expedited notifications

Section titled “Monitored impressions, manual classifications, expedited notifications”

The three usage counters of a plan, per calendar month (UTC):

  • Monitored impressions: ad impressions that AD80 observed and processed through your SDK. This counts impressions, not unique creatives.
  • Manual classifications: images you upload or pick as samples in the Policy Playground, and re-tests.
  • Expedited notifications: alerts sent right away. When the monthly allowance is used, later alerts arrive by digest.