Changelog
API and product changes for developer integrations.
16 November 2026
Postman collection for the developer API
- New Test with Postman page with downloadable collection and Sandbox / Production environment files covering every documented developer endpoint (verify, sessions, extract, screening, monitoring, webhooks, health, screening PDF).
- CI validates the collection stays on Postman v2.1 and matches the documented route list so it cannot drift from the API reference.
- Postman docs and collection description call out selecting the Sandbox or Production environment and setting
apiKey(fixes401 Missing API keywhen No Environment is selected). - Document extract request in the collection now uses
drivers_licencewith requireddocumentBackImageplaceholders.
Document extract documentation
- Document extract documents which
documentTypevalues requiredocumentBackImage(onlydrivers_licence), the image object shape, driver licence front and back examples, and error code tables for extract and related verify routes includingDOCUMENT_BACK_REQUIRED. - Request size limits: 6,291,456 byte API entry body cap (413 before the application), 5,000,000 bytes decoded per image, resizing guidance for two-image licence extract, and Postman/MCP updates aligned with those limits.
Screening always includes adverse media
- Every live screening search now runs sanctions and watchlists, PEP, and adverse media together. The half-credit screening charge is unchanged.
listFilter.includeAdverseMediaremains accepted for backward compatibility. When set tofalse, it is ignored.- Screening PDF reports include dedicated sanctions, PEP, and adverse media sections. When a data point has no hits, the PDF states that it was searched with the search date and time (AEST).
14 November 2026
Form links: document scan autofill
- After ID photos are accepted on a biometric form link, the page reads empty fields from the same document extract pipeline as
POST /v1/verify/document-extract(free, no credits). Values from your APIprefillare never overwritten. Scan-filled fields stay editable even whenallowEditingPrefilledFieldsisfalse. - Document and back-of-licence capture complete responses may include
fieldSuggestionsfor the browser to apply locally.
Form links: selfie step completion
- Fixed a Safari issue where the selfie check could show a blank screen after capture by keeping the liveness widget mounted until analysis finishes, matching hosted verify behaviour.
Prefilled form links: lock prefilled fields by default
- Embedded form configurations include
allowEditingPrefilledFields(dashboard toggle, defaultfalse). When off, non-emptyprefillvalues onPOST /v1/forms/:formId/sessionsare shown read-only on the public form and enforced at submit. Empty prefill values stay editable. - Form link session responses include
lockedPrefillFieldKeyswhen editing is disallowed.
Hosted and form-link camera capture retake
- SMS hosted verify and prefilled form links (embedded capture) now ask the subject to retake a document photo once when quality checks fail, then let them continue and record flags on the session result. Driver licence back capture uses the same rule for the back photo. Passport and other portrait IDs use the rule on the single document photo.
- Session GET responses for hosted and form-link subjects include
captureProgress(licence side and retake prompts) so a refreshed page can resume the correct step. - Capture presign errors use specific messages (missing front, finish front upload, or front retake required) instead of a generic front-first error.
- Embedded form branding applies to in-flow camera buttons and guides. When only
primaryis set,primaryForegroundis derived for accessible contrast on primary buttons. - Document images remain downloadable for 24 hours after capture via
GET /v1/verify/sessions/:id/captures/documentand…/document_back, including when quality flags are present.
Screening outcomes aligned across dashboard, API, PDF, and webhooks
- Screening responses and monitoring webhooks now include
outcomeandoutcomeLabel(Clear,Possible match, orPotential match) using the same rules as the dashboard and PDF.clearedis unchanged (truewhen no match scored ≥85). - Screening PDFs list every stored hit, including possible matches below the strong match threshold, in separate sections.
- Removed optional match fields
sourceUrl,sourceDataset,externalId, andlastSeenfrom screening API responses, monitoring webhooks, and the dashboard hit view. UsesourceNamefor the list display title.
9 November 2026
Document extract quality
verifyRequestdate fields are normalised toYYYY-MM-DD(orYYYY-MMfor Medicare green card expiry). Unparseable dates are omitted with a flag instead of invalid strings.- Passport responses without a readable MRZ now include a visual-zone fallback (still flagged) for surname, given names, passport number, and date of birth when found on the data page.
- Label bleed filtering reduces card titles and field labels being returned as person or number fields.
Document extract API
- New
POST /v1/verify/document-extractreads photos for all twelve verification document types and returns structured fields plus averifyRequestobject forPOST /v1/verify. Capture and extraction concerns are returned as flags only. - Uploaded images can be retrieved for 24 hours via
GET /v1/verify/document-extract/:id/captures/documentand…/document_back. - See the new Document extract docs page for request limits and supported document types.
- Document extract is free and does not use verification credits on test or live API keys.
- Test MCP server: new
extract_documenttool (same behaviour as document extract, no credits).
Screening PDF reports (API and dashboard)
- New endpoint
GET /v1/screening/:id/reportreturns a brandedapplication/pdfscreening summary. Scoped to your API key account and environment.GET /v1/screening/:idand livePOST /v1/screeningresponses includereportUrl. - Identity verification PDFs are available from the customer dashboard only (not on the public API). Use Download PDF on identity checks and the developer usage log.
- Screening PDFs list subject details, overall Clear or Potential match, and each list hit separately with plain-English labels.
- Test MCP server:
screen_subjectandget_screening_requestincludereportUrl; newget_screening_reportfetches the PDF (and metadata) for a screening id.
7 November 2026
Hosted verify: non-blocking capture checks and 24 hour document images
- Document photo checks (quality, authenticity, portrait detection) no longer stop the subject on the phone. Concerns are recorded in
result.captureCheckFlagsandresult.documentCaptureChecks. The UI may show a short optional retake tip; the subject can always continue. - Text cross-checks (entered fields vs photo, including passport MRZ validation) no longer block the hosted flow. Mismatches appear as
document_text:*flags on the session result and webhooks. - Face match no longer fails the session when the compare step is low. Use
faceMatchScoreandface_match:FACE_MATCH_FAILEDin your own rules. - New endpoints:
GET /v1/verify/sessions/:id/captures/documentandGET /v1/verify/sessions/:id/captures/document_backreturn a short-liveddownloadUrlfor the stored JPEG (24 hours after capture). After that window the API returns HTTP410witherror: "EXPIRED". - Document capture images are retained in hosted Australian data storage for 24 hours, then deleted. Liveness reference frames are still removed soon after processing.
Details: Hosted verify sessions, Capture check flags.