generated: '2026-07-21'
method: derived
source: >-
  Derived from the schemas in openapi/voiceitt-rest-api-openapi-original.json
  (SignInUserIdDTO, SignInEmailDTO, RefreshTokenDTO, AuthResponseDTO,
  TranscribeResponseDTO) and the id-reference fields they carry (app_id,
  api_key, user_id), plus the session/streaming model documented at
  https://voiceitt-si-api.readme.io/.
docs: https://voiceitt-si-api.readme.io/
notation: >-
  relationships use has_one / has_many / belongs_to with the foreign-key field
  name; direction is from the entity that owns the reference.
entities:
- name: Application
  description: A developer app registered in the Voiceitt developer portal, identified by app_id and authenticated with api_key.
- name: User
  description: An end user — anonymous (user_id, max 36 chars, speaker-independent) or enrolled with email/password for a personalized speech model trained at web.voiceitt.com.
- name: AuthSession
  description: A JWT session — token + refresh_token with expiry timestamps (token_expires_at, refresh_token_expires_at).
- name: SpeechModel
  description: The recognition model backing a session — speaker-independent, or personalized per enrolled user (WebSocket lifecycle events model_loading / model_ready / model_missing).
- name: Transcription
  description: A recognition result — HTTP transcribe returns text; streaming returns recognition / partial_recognition segments with segment_id, start_time, end_time.
relationships:
- {from: User, to: Application, kind: belongs_to, via: app_id}
- {from: AuthSession, to: Application, kind: belongs_to, via: app_id}
- {from: AuthSession, to: User, kind: belongs_to, via: user_id}
- {from: Application, to: AuthSession, kind: has_many, via: app_id}
- {from: User, to: SpeechModel, kind: has_one, via: user_id}
- {from: AuthSession, to: Transcription, kind: has_many, via: token}
- {from: Transcription, to: User, kind: belongs_to, via: user_id}