Créer une transcription

curl --request POST \ --url https://api.dgrid.ai/v1/audio/transcriptions \ --header 'Authorization: Bearer <token>' \ --header 'Content-Type: multipart/form-data' \ --form file='@example-file' \ --form 'model=<string>' \ --form 'language=<string>' \ --form 'prompt=<string>' \ --form response_format=json \ --form temperature=0

Autorisations

Authorization

string

header

requis

Your DGrid API key. All endpoints use Authorization: Bearer <DGRID_API_KEY>.

Corps

multipart/form-data

file

requis

Audio file up to 25 MB.

model

string

requis

Model ID, such as whisper-1.

language

string

ISO-639-1 language code, such as en, zh, or ko.

prompt

string

Optional prompt for biasing the transcript.

response_format

enum<string>

défaut:json

Options disponibles:

json,

text,

srt,

verbose_json,

vtt

temperature

number

défaut:0

Sampling temperature.

Plage requise: 0 <= x <= 1

Réponse

Transcript. When response_format is verbose_json, the response also includes task, language, duration, and per-segment timing metadata.

text

string

Transcript text.