local-transcribe

name	local-transcribe
description	Transcribe audio files (m4a, mp3, wav, ogg, flac, aac, webm, mp4) to text using local whisper-cpp (offline, no API). Use when asked to transcribe locally, convert speech to text offline, or process audio recordings without sending data externally. Converts to 16kHz WAV via ffmpeg, then runs whisper-cli locally. Does not do speaker diarization or advanced transcription.
argument-hint	<audio-file> [--output PATH] [--model MODEL_PATH]
allowed-tools	Read, Write, Edit, Glob, Bash(skills/local-transcribe/scripts/transcribe.sh *), Bash(chmod +x *)

Audio Transcription

Transcribe audio files to text using whisper-cpp (local, offline).

Script location

skills/local-transcribe/scripts/transcribe.sh

Basic usage

skills/local-transcribe/scripts/transcribe.sh "<audio-file>"

The transcript is printed to stdout. Use --output to save to a file instead.

Options

• --output PATH — save transcript to a file instead of stdout
• --model PATH — path to a whisper.cpp GGML model file (auto-detected if omitted)

Examples

# Transcribe and print to stdout
skills/local-transcribe/scripts/transcribe.sh recording.m4a

# Transcribe and save to file
skills/local-transcribe/scripts/transcribe.sh recording.m4a --output transcript.txt

# Use a specific model
skills/local-transcribe/scripts/transcribe.sh recording.m4a --model ~/.cache/whisper-cpp/ggml-medium.bin

Supported formats

Any format ffmpeg can decode: .m4a, .mp3, .wav, .ogg, .flac, .aac, .webm, .mp4

Dependencies

• whisper-cpp — the transcription engine
  • macOS: brew install whisper-cpp
  • Linux (apt): sudo apt-get install whisper.cpp
  • Linux (arch): yay -S whisper.cpp-git
  • From source: https://github.com/ggerganov/whisper.cpp
• ffmpeg — for audio format conversion
  • macOS: brew install ffmpeg
  • Linux: sudo apt-get install ffmpeg
• A GGML model file — auto-detected in ~/.cache/whisper-cpp/, or specify with --model
  • Download: curl -L -o ~/.cache/whisper-cpp/ggml-small.bin https://huggingface.co/ggerganov/whisper.cpp/resolve/main/ggml-small.bin

Models

Model	Size	Speed	Quality
ggml-tiny.bin	75 MB	Fastest	Lower accuracy
ggml-base.bin	142 MB	Fast	Decent
ggml-small.bin	466 MB	Moderate	Good (recommended)
ggml-medium.bin	1.5 GB	Slower	Better
ggml-large.bin	3.1 GB	Slowest	Best

The script searches ~/.cache/whisper-cpp/ for models automatically, preferring small > base > medium > tiny.

After transcribing

The raw transcript is plain text. You can then:

• Clean it up (remove filler words, fix punctuation)
• Add it to a markdown note
• Use it as input for further processing