openai-whisper-api

You can transcribe audio files into text using OpenAI Whisper by running a simple bash script that posts your audio file to the Whisper API endpoint via curl, returning a text or JSON transcript.

Yes, you can automate audio transcription in bash scripts for batch processing. This curl-based workflow is designed to be automation-friendly, making it ideal for processing meeting minutes and media archives in automated pipelines.

Yes, you need an OpenAI API key and curl to use this Whisper transcription workflow. The script requires curl installed on your system and an OPENAI_API_KEY configured to authenticate the API calls.

The best way to configure Whisper transcription outputs is by specifying the desired model, language, and output format. The script defaults to the whisper-1 model and outputs a .txt file next to your input audio.

Your curl audio transcription script might not return JSON output because the default output format is text. You need to explicitly configure the script to request the JSON output format to receive structured transcript data.