TsgcHTTP_API_OpenAI › Methods › CreateTranscriptionFromFile
Transcribes a local audio file into text using an OpenAI speech-to-text model
function CreateTranscriptionFromFile(const aRequest : TsgcOpenAIClass_Request_Transcription) : TsgcOpenAIClass_Response_Transcription;| Name | Type | Description |
|---|---|---|
aRequest | const TsgcOpenAIClass_Request_Transcription | Transcription request where FileName points to a local audio file on disk |
Transcription response containing the recognized text and optional metadata (TsgcOpenAIClass_Response_Transcription)
Calls the POST /v1/audio/transcriptions endpoint with a multipart/form-data body that uploads the local audio file directly. Use this overload when you have the recording stored locally. Supported formats include mp3, wav, m4a, mp4, mpeg, mpga and webm, with a 25MB file size limit.
oRequest := TsgcOpenAIClass_Request_Transcription.Create;
oRequest.Model := 'whisper-1';
oRequest.FileName := 'C:\audio\meeting.mp3';
oRequest.Language := 'en';
oResponse := oAPI.CreateTranscriptionFromFile(oRequest);
ShowMessage(oResponse.Text);