generated: '2026-08-28'
method: searched
source: https://sondehealth.atlassian.net/wiki/spaces/SA/pages/2706866198/API+Release+Notes+and+Best+Practices
docs: https://sondehealth.atlassian.net/wiki/spaces/SA/pages/2706866198/API+Release+Notes+and+Best+Practices
scheme: numbered releases (major.minor), dated, published as a Confluence page
current_version: '3.4'
current_version_date: '2022-08-29'
staleness_note: >-
  The newest dated entry on Sonde's API release-notes page is API 3.4 (29 August 2022).
  No cloud-API release has been published in the four years since; Sonde's more recent
  developer activity in the same documentation space is on the Passive/Edge mobile SDKs,
  which have no dated changelog of their own.
entry_count: 7
entries:
- version: '3.4'
  date: '2022-08-29'
  breaking: false
  highlights:
  - M3 questionnaire deployment and scoring
  - Partner questionnaire creation
  additions:
  - POST /platform/v2/inference/inference-parameters
  - POST /platform/v2/inference/scores
  - GET /platform/v2/inference/scores/{scoreId}
  - POST /platform/v1/questionnaire
- version: '3.3'
  date: '2021-09-20'
  breaking: false
  highlights:
  - Crispness acoustic score added to Mental Fitness Voice Features
  - Aggregate Mental Fitness score (integration of the 6 acoustic scores) added
  changes:
  - GET /platform/async/v1/inference/voice-feature-scores/{asyncJobId}
- version: '3.2'
  date: '2021-09-06'
  breaking: false
  highlights:
  - Inference v2 APIs introduced to infer a measure score from two audio samples
  additions:
  - POST /platform/v2/inference/inference-parameters
  - PUT /platform/v2/inference/inference-parameters/{inferenceParameterId}
  - POST /platform/v2/inference/scores
  - GET /platform/v2/inference/scores/{scoreId}
- version: '3.1'
  date: '2021-06-16'
  breaking: false
  highlights:
  - TranscriptionService — speech-to-text for health-condition audio files
  - Voice-feature-score history filterable by time (from/to) and user
  additions:
  - POST /platform/async/v1/transcriptions
  - GET /platform/async/v1/transcriptions/{asyncJobId}
  - GET /platform/v1/storage/files/{filePath}/transcriptions
  - GET /platform/v1/inference/voice-feature-scores
  - GET /platform/v1/inference/voice-feature-scores/{voiceFeatureScoreId}
- version: '3.0'
  date: '2021-05-24'
  breaking: false
  highlights:
  - Acoustic voice features for Mental Fitness
  additions:
  - POST /platform/v2/users
  - POST /platform/async/v1/inference/voice-feature-scores
  - GET /platform/async/v1/inference/voice-feature-scores/{asyncJobId}
  - GET /platform/v1/measures/name/{measureName}/voice-features
- version: '2.6'
  date: '2021-04-07'
  breaking: false
  behaviour_change: true
  highlights:
  - >-
    Elicitation Check (ELCK) for Respiratory Symptom Risk voice samples — automatic
    validation of energy level, voice presence, and correct elicitation (a single
    continuous "ahh" held for 6 seconds). Failing samples are rejected.
  - >-
    New failure response on POST /platform/v1/inference/scores — HTTP 422 with error
    code ELICITATION_CHECK_FAILED. No additional API calls are required to opt in,
    so integrators saw a new rejection path without changing code.
  notes:
  - Sonde states Elicitation Check accuracy is approximately 80%.
  - Sonde recommends asking for no more than 3 voice samples before treating the test
    as unavailable.
- version: '2.5'
  date: '2021-02'
  breaking: false
  highlights:
  - Backward compatible with API 1.0, adding the API 2.0 and 2.5 resources
  additions:
  - QuestionnaireManager — return a questionnaire to show in the partner app and capture
    its responses
  - UserService — update device type and device manufacturer
  - InferenceService — historical score data for one or many subjects over a date/time
    range
  - StorageService — download submitted audio files
deprecations_ref: lifecycle/sonde-health-lifecycle.yml