Skip to main content
Version: 1.25

Speech to Text (Whisper)

Audience: Citizen Developers

Skill Prerequisites: Tokens

This action extracts text from an audio file using the OpenAI Whisper API.

Typical Use Cases

  • Convert speech to text in an audio file
  • Transcribe audio recordings into written content

Don't use it to

  • Extract text from non-audio files
  • Transcribe audio content not supported by OpenAI Whisper

None

Input Parameter Reference

ParameterDescriptionSupports TokensDefaultRequired
OpenAi API KeyThe API key for Open AI.Yesempty stringYes
Audio fileIdentifier for the audio file. Can be: RelativeUrl, AbsoluteUrl, LinkClickUrl, or PhysicalPath.Yesempty stringYes
Response FormatChoose from the following: json, text, srt, verbose_json, vtt.YesjsonNo
PromptOptionally, provide additional instructions for OpenAi to use while creating the transcript.Yesempty stringNo

Output Parameters Reference

ParameterDescription
Output Token NameStores the extracted text in a token with the specified name.

Examples

1. Basic Example

Extract text from an audio file and store the result in a token named "Transcript".

[
{
"Id": -1,
"ActionErrorMessage": null,
"ActionType": "OpenAi.Whisper",
"Condition": null,
"Description": null,
"IsDeleted": false,
"EventName": "",
"Parameters": {
"ApiKey": "your_openai_api_key",
"AudioFileIdentifier": "path/to/your/audio/file.mp3",
"ResponseFormat": {
"IsExpression": false,
"Expression": null,
"Value": "text"
},
"Prompt": null,
"OutputTokenName": "Transcript"
}
}
]

For example, you can import this JSON content into a workflow using the "Import Actions" button.