Transcribe

by badlogic90bb51cae36515a648515b633a81c0c6efc8c74dNo license2.5K starsListed Oct 9, 2026Updated Oct 9, 2026Repository updated 4 months ago

Local speech-to-text transcription on Apple Silicon macOS. Supports wav directly and other audio formats via ffmpeg.

AI-generated overview

Transcribes audio to timestamped text locally on Apple Silicon macOS using parakeet-cpp-transcribe.

What it does: Runs a shell script that invokes parakeet-cpp-transcribe for local speech-to-text on Apple Silicon macOS. It accepts WAV directly and other audio formats through ffmpeg, and writes plain text timestamped in 15-second chunks to stdout. On first run it downloads the macOS arm64 binary from a GitHub release into an ignored bin directory, and the binary fetches its GGUF model automatically if missing. Diagnostic model logs go to stderr.
When to use it: Use it when you need speech-to-text transcription of an audio file on an Apple Silicon Mac and want the processing to happen locally. It suits producing timestamped plain-text transcripts, including from non-WAV formats when ffmpeg is installed.
Requirements: Apple Silicon macOS only; curl and tar; ffmpeg for non-WAV input (brew install ffmpeg); network access to download the binary from a GitHub release and the GGUF model; ships an executable script (transcribe.sh).

Local speech-to-text using parakeet-cpp-transcribe on Apple Silicon macOS.

The first run downloads the macOS arm64 binary from the latest badlogic/pibot GitHub release into the extension's ignored bin directory:

({extensionDir} is the parent directory of this skill's {baseDir}.) The binary downloads its GGUF model automatically if missing.

Plain text timestamped in 15 second chunks is written to stdout:

Model/GGML diagnostic logs are written to stderr. Redirect stderr to hide them:

Source and attribution

License: No license

Content belongs to its original authors. SourceWeft indexes it from a public repository.