mh-openai-whisper
by mohdalhashemi98-hue
AI & LLMsBrief
Local speech-to-text with the Whisper CLI (no API key).
Description
What This Skill Does
Runs OpenAI's Whisper speech recognition model locally via the command line. Transcribes or translates audio files to text without sending data to any external API. Models are cached locally after the first download.
Runs entirely on-device, so audio never leaves your machine and there are no per-request API costs.
When to Use It
- Transcribing recorded meeting audio to a text file
- Generating SRT subtitle files for video content
- Translating foreign-language audio directly to English text
- Batch transcribing voice memos offline
- Converting podcast episodes to searchable text
Example Workflow
User asks: transcribe this interview recording to a text file
- 1
Identify the audio file path provided by the user
- 2
Run whisper with --model medium, --output_format txt, and --output_dir set to current directory
- 3
Wait for the model to download if not already cached
- 4
Return the path to the generated .txt file
interview.txt containing the full transcript
Installation
clawhub install mohdalhashemi98-hue/mh-openai-whisperRun the CLI command, or paste the prompt into Claude Code, Cursor, Codex or whichever coding agent you use.