October 2026
Risk signal names
The current scored webhook usesprivate_relay, network_check_incomplete, latency_test and banned_ip for Private Relay, Network Check Incomplete, Latency test and Banned IP. The connection type privacy_relay is unchanged.
Plan request rate
Each registered domain now allows fewer identification requests per second: 3 on Free and Starter, 5 on Growth, and 7 on Scale. See Billing and Rate limits.Current product and integration reference
- History responses, scored webhooks and the server SDKs name one check
identification_id. - Named the glossary field for an identification’s Trusted, Suspicious, or Dangerous label Risk band.
- Grouped the API reference sidebar by the integration steps: run a check, get the result, account, and contracts.
- Corrected risk weights, including VPN 30, Proxy 30, Private Relay 30, IP Mismatch 45 and Browser Automation 90.
- Distinguished source detections from grouped dashboard labels.
- Documented the separate multi-account group webhook, its enablement, lifecycle and revisions.
- Updated webhook retry and deduplication guidance and separated History fields from scored payloads.
- Corrected current identity risk and the Traffic quality.
- Replaced dashboard illustrations with screenshots of the current analytics dashboard code, using synthetic example data, including the account-event graph.
- Shortened use-case guides and linked them to complete runnable tutorials.
- Updated the scored webhook flags and signal catalog for the 20-field schema.
- Split the glossary into entities, the fields of an identification, risk signals and a separate integration glossary.
- Session ID and Cookie ID are identifiers on a check, not entities. The identifier model separates the identification, the user and the visitor from the device anchor.