openai-whisper-api

• You have audio files (e.g. .m4a, .ogg) you need transcribed quickly without building custom API integration code.
• You want to automate transcription as part of a shell pipeline or scripted workflow.
• You need to pass a language hint or speaker-name prompt to improve transcription accuracy.
• You want structured JSON output from the Whisper API rather than plain text.

• You need real-time or streaming transcription — this skill sends complete audio files to the API and is not suitable for live audio.
• You are not comfortable setting an OPENAI_API_KEY, as the skill has no alternative transcription backend.

Run a single shell script with a path to your audio file and receive a .txt transcript alongside it by default. No boilerplate API code required.

Write output to a custom path with --out, or request raw JSON from the API with --json for downstream processing.

Use --language to specify the spoken language (e.g. en) for better accuracy, or --prompt to supply context like speaker names that help guide the transcription.

Defaults to whisper-1 but accepts a --model flag, keeping you ready for future Whisper model variants offered by OpenAI.

Record a meeting or interview, then run the skill against the audio file to produce a full text transcript. Pass --prompt with participant names to improve speaker-name recognition in the output.

Integrate the script into a cron job or CI workflow that picks up new audio files from a directory and transcribes them automatically, writing results to a specified output folder.

Convert a library of voice memos or podcast episodes into searchable text files by scripting the skill over a batch of audio files.