Skip to content

audio_transcription.md

Latest commit

 

History

History
93 lines (69 loc) · 2.78 KB

audio_transcription.md

File metadata and controls

93 lines (69 loc) · 2.78 KB
Raw
page_title openai_audio_transcription Resource - terraform-provider-openai
subcategory
description Creates an audio transcription. Note: This resource does not support updates - any configuration change will create a new resource.

openai_audio_transcription (Resource)

Creates an audio transcription. Note: This resource does not support updates - any configuration change will create a new resource.

Example Usage

resource "openai_audio_transcription" "example" {
  file  = "./speech.mp3"
  model = "whisper-1"

  # Optional parameters
  language        = "en"
  prompt          = "This is a transcription of a meeting"
  response_format = "json"
  temperature     = 0
}

# Example with verbose output
resource "openai_audio_transcription" "verbose_example" {
  file  = "./speech.mp3"
  model = "whisper-1"

  response_format = "verbose_json"
  temperature     = 0.2

  # Get word-level timestamps (only works with verbose_json)
  timestamp_granularities = ["word", "segment"]
}

output "transcription_text" {
  value = openai_audio_transcription.example.text
}

Schema

Required

  • file (String) Path to the audio file to transcribe.
  • model (String) ID of the model to use (e.g., 'whisper-1', 'gpt-4o-transcribe').

Optional

  • include (List of String) Additional information to include (e.g. 'logprobs').
  • language (String) The language of the input audio (ISO-639-1 format).
  • prompt (String) An optional text to guide the model's style or continue a previous audio segment.
  • response_format (String) The format of the transcript output (e.g. json, text, srt).
  • temperature (Number) The sampling temperature, between 0 and 1.

Read-Only

  • duration (Number) Duration of the audio in seconds.
  • id (String) The identifier of the transcription (randomly generated).
  • segments (List of Object) Segments of the transcription. (see below for nested schema)
  • text (String) The transcribed text.

Nested Schema for segments

Read-Only:

  • end (Number)
  • id (Number)
  • start (Number)
  • text (String)

Import

Import is supported using the following syntax:

The terraform import command can be used, for example:

#!/bin/bash
# Audio transcription resources cannot be imported

echo "Audio transcription is a one-time operation, not a persistent resource."
echo "It cannot be imported because there's no state to track in OpenAI."
echo ""
echo "To use audio transcription:"
echo "1. Create a resource configuration (see resource.tf)"
echo "2. Run 'terraform apply' to perform the transcription"