IMimentiv
Audio Emotion API

Get Audio Translation

Changed on

Read one stored translation.

Access is the same as for the list. A translation that is not ready is not found for a caller who is not the audio's owner.

get/v2/audios/{audio_id}/translations/{language}

Request

  • Base URL
    https://api.imentiv.ai (the document declares no server; this is the origin it was published from)
  • Auth
    • API key in header X-API-Key
    • or HTTP bearer
    • or OAuth 2
      Authorization code: https://api.imentiv.ai/oauth/authorize
      Token URL: https://api.imentiv.ai/oauth/token scopes: media.read

Path parameters

audio_idstring required

The unique ID of the audio.

languagestring required

The target language as a BCP-47 tag, for example es or pt-BR.

Response

Successful Response

languagestring required

The target language, as a canonical BCP-47 tag.

status'queued' | 'running' | 'ready' | 'failed' required

queued, running, ready or failed.

linesstring[] nullable

One entry per transcript segment, in the transcript's order. Present only when status is ready. A segment with no text comes back as an empty string. A segment that was already in the target language was not translated and comes back as null: show its own text.

source_revisionstring required

The transcript revision this translation was requested for.

source_line_hashesstring[] nullable

For each segment, the SHA-256 (64 lowercase hexadecimal characters) of the segment's text (trimmed with ECMAScript trim, then NFC-normalised) that the translation was made from; null for a segment that was not translated. Compare with the current segment to tell which line is out of date. Null for a translation made before these were recorded: use stale.

stale_linesinteger[] nullable

The indexes of the segments whose text changed since this translation was made, when it can be said line by line: the number of segments and each segment's time and speaker are unchanged. Empty while stale is false. Null when stale is true and it cannot be said line by line: treat the whole translation as out of date.

staleboolean required

True once the transcript's text or timing has changed since this translation was made. Renaming, merging or deleting speakers does not change it.

error_codestring nullable

Why a failed translation failed. Present only for the owner.

attempt_charge_state'reserved' | 'captured' | 'released' nullable

reserved while an attempt can still be charged, then captured if it was charged or released if it was not. Present only for the owner.

captured_creditsinteger nullable

Credits charged for the attempt. Present only for the owner.

reserved_creditsinteger nullable

Credits held for the attempt that have not been settled yet; 0 once settled. Present only for the owner.

estimated_creditsinteger nullable

The estimate this attempt reserved. The charge is the actual usage and can differ. Present only for the owner.

retry_kind'resend' | 'new_attempt' nullable

For a failed translation, what POST .../retry must send: resend repeats the original request, new_attempt needs the current transcript revision and a new idempotency_key. Present only for the owner.

Changes