Skip to content

Repository files navigation

transcribe

Local, offline transcription for macOS 14+ on Apple Silicon. Turn audio/video or a YouTube URL into a readable transcript with optional speaker diarization.

Install

uv tool install git+https://github.com/speech115/transcribe.git

Requires ffmpeg; install yt-dlp for YouTube sources.

Verify the installed executable and print the bundled agent skill with:

transcribe doctor
transcribe skill

Quick start

transcribe call.m4a --speakers auto

The command prepares audio, runs vendored FluidAudio/Parakeet, optionally diarizes, merges turns, and writes transcript.md, transcript.json, and manifest.json (plus requested SRT/VTT subtitles).

Streaming diarization runs alongside ASR. Offline diarization runs after ASR.

Language and speaker count are automatic by default. Use --lang, --speakers, --out, --out-root, and --formats when needed.

SRT/VTT cues use up to 80 text characters and 6 seconds; a single word keeps its original timing. Existing transcript artifacts require --overwrite. Artifacts are prepared in a temporary subdirectory of the output before publication, with rollback on failure. The output directory and unrelated files are left in place. A forced process kill during publication can leave a partial result; previous artifacts remain in the output's .transcribe-*/previous directory.

Privacy

Processing is local after the model cache is present. Media is not uploaded; network access is used only for the initial model cache and YouTube download.

Status

v0.5.0 · single-maintainer project · Apache-2.0

About

Local, offline audio/video transcription with speaker diarization for Apple Silicon.

Resources

Contributing

Security policy

Stars

2 stars

Watchers

0 watching

Forks

Releases

Packages

Contributors

Languages