Integrations Overview

AppRefer forwards conversion events to ad networks via server-side Conversions APIs. No client-side pixels or SDKs needed on the ad network side.

Supported Integrations#

IntegrationTypeDescription
Meta (Facebook)Conversions API (CAPI)Server-side event forwarding using fbclid. Supports Advanced Matching.
Google AdsOffline ConversionsClick conversion uploads via gclid. Auto token refresh.
TikTokEvents APIServer-side event forwarding using ttclid.
RevenueCatWebhooksReceive all purchase, subscription, and billing events. No in-app tracking code needed.
SKAdNetworkApple AttributionPrivacy-safe iOS attribution with SKAN 4.0 conversion values.

How Forwarding Works#

When a user triggers an event (install, purchase, trial start, etc.), AppRefer follows this flow:

  1. Attribution is resolved -- the device is matched to a click with a click ID (fbclid, gclid, or ttclid).
  2. Event occurs -- an install, purchase, trial start, or custom event is recorded.
  3. Integration config lookup -- the server checks which ad networks are configured for this app.
  4. Credentials loaded -- stored access tokens and IDs are retrieved securely.
  5. Event forwarded -- the event is sent to the ad network's server-side API with the original click ID.
  6. Result logged -- success or failure is recorded in the audit log.
  7. Retry on failure -- failed forwards are retried automatically with exponential backoff.

Sandbox Isolation

Events from pk_test_ keys are never forwarded to ad networks. Sandbox events are logged for debugging but completely isolated from production forwarding.

Test Connection#

Before going live, verify your credentials are correct. In the dashboard, go to Settings → Integrations, select a network, and click "Test Connection". This sends a dry-run request to the ad network API to confirm your credentials are valid without recording any actual conversion.

Getting Started

Choose an integration above to see detailed setup instructions, event mapping tables, and configuration requirements for each ad network.