Skip to content
AmplifierDocs
Esc
navigateopen⌘Jpreview
On this page

Production Checklist

Ship a production integration against the Voice Biomarker API.

Audio quality

  • Tested recording setup with sample audio in the target environment; audio_quality.issues is empty in test runs
  • Production recording pipeline uses 16 kHz mono WAV or FLAC (recommended; the API also accepts MP3, M4A, and 8 kHz minimum)
  • Single-speaker capture confirmed
  • Microphone placement follows the guidelines in Audio Requirements

Credential management

  • X-Account-ID and X-API-Key are stored as environment variables or in a secrets manager, rather than in source code
  • Credentials are excluded from all logging pipelines
  • Access to credentials is restricted to the services that need them

Error handling

  • 401 (authentication failure) surfaces an alert to the integration team
  • 402 (no credits) surfaces a clear message and triggers a support notification
  • 400 / 422 errors (AUDIO_TOO_SHORT, AUDIO_TOO_LONG, UNSUPPORTED_FORMAT, AUDIO_POOR_QUALITY) return a user-facing prompt to re-record or correct the request
  • 429 (rate limit) handled with exponential backoff; Retry-After header is read
  • 500 errors retried with exponential backoff (max 5 retries)
  • See Errors for the complete error code list and retry guidance

Display rules

  • level and label are used as the display primitives in user-facing views; score values are kept for internal logging and analytics
  • summary.description is shown only after review by qualified staff, and stays out of automated alerts and employee- or patient-facing output
  • flagged: true is used to filter which signals appear in user-facing views
  • Display labels follow the recommended mapping in Display guidelines

Analysis selection

  • If using the Model API, the model_name matches the population — pulse for general wellness unless a specialized bundle applies
  • If using the Sign API, the sign_name targets the specific signal relevant to your workflow
  • If your model or sign includes Emerging or Investigational signs, appropriate clinical oversight is in place
  • Your team understands how per-sign evidence tiers relate to your workflows (see Evidence tiers and Choosing a Model or Sign)

Data handling

  • Data handling and retention requirements reviewed with your legal and compliance team
  • For employee wellness deployments: individual employee results have appropriate privacy controls, and aggregated team-level signals (not individual scores) are used for dashboards and manager views

Webhooks (if applicable)

For how delivery, retries, and signature verification work, see Webhooks.

  • Webhook URL is HTTPS
  • Receiver inspects status on the payload (done, failed, or timed-out) before reading result
  • Signature verification is implemented before processing any webhook payload — see Signature verification
  • Idempotency check on job_id is implemented to handle webhook retries
  • Webhook receiver responds within 30 seconds to stay inside the delivery timeout

Final verification

  • End-to-end test completed with a real recording in the target environment
  • audio_quality.issues verified as empty in the production environment test
  • escalate workflow tested with a simulated high-signal result
  • consider workflow tested to verify signals are surfaced and routed appropriately
  • recommended_action of none or inconclusive cases handled gracefully (no alert, log only)

Was this page helpful?