How to send audio file to Whisper API

API
1

Hello,

I need some help I’m using the following code in php to translate an audio to text, but I always get the same error.
{
“success”: false,
“message”: “An error occurred: A ‘contents’ key is required”
}

When I add key contents it tells me that the file field is missing

// Make API request to transcribe audio
$response = $client->post(‘https://api.openai.com/v1/audio/transcriptions’, [
‘headers’ => [
‘Authorization’ => “Bearer {$apiKey}”,
‘Content-Type’ => ‘application/json’,
],
‘multipart’ => [
[
‘name’ => ‘audio_file’,
‘file’ => fopen($audioPath, ‘r’),
]
],
‘json’ => [
‘model’ => ‘whisper-1’
]
]);

2

Content-Type is not “JSON”.

Example from API dox:

curl https://api.openai.com/v1/audio/transcriptions \
  -H "Authorization: Bearer $OPENAI_API_KEY" \
  -H "Content-Type: multipart/form-data" \
  -F file="@/path/to/file/audio.mp3" \
  -F model="whisper-1"
3

Hello,

Thanks for your answer/help, I changed the content-type and now I get the answer

{
“success”: false,
“message”: “An error occurred: json_encode error: Type is not supported”
}

I changed the code to:

// Make API request to transcribe audio
$response = $client->post(‘https://api.openai.com/v1/audio/transcriptions’, [
‘headers’ => [
‘Authorization’ => “Bearer {$apiKey}”,
‘Content-Type’ => ‘multipart/form-data’,
],
‘json’ => [
‘model’ => ‘whisper-1’,
‘name’ => ‘audio_file’,
‘file’ => fopen($audioPath, ‘r’),
]
]);

4

I’ve already managed to run the audio transcription, but the transcription model is unsatisfactory.

Thanks for the response

5

You can read about whisper prompting, to improve the interpretation of the audio with not just a previous transcript to continue on, but also made up prompts to influence the audio interpretation.