A simple, on-device transcription utility for your audio files.
Slap an audio file in, get a text file out.
Jotty is a no-frills command-line tool that uses Apple's Speech framework to transcribe your audio files. It's designed to be simple and work entirely on your machine.
It doesn't spew text into your console. Instead, it saves a transcript file, adding .jotty.txt or .jotty.json to your original file's name.
- macOS 26.0 (Tahoe) or later
(wip)
$ brew tap ys319/homebrew-tap
$ brew install jottyDownload the latest binary from the Releases page.
Just point it at an audio file.
# Get a plain text transcript
$ jotty ./meeting.mov
# Get a JSON output instead
$ jotty ./meeting.mov --format json
# Overwrite the old transcript if it exists
$ jotty ./meeting.mov --overwrite
# Transcribe a file in a specific language (e.g., Japanese)
$ jotty ./meeting.mov --language ja-JPFor all the details, here's the --help output:
OVERVIEW: On-device transcription utility for audio files.
Uses Apple's Speech framework to transcribe the provided audio file.
Required speech models are verified automatically and downloaded when needed.
USAGE: jotty <input-file-path> [--format <format>] [--language <language>] [--overwrite]
ARGUMENTS:
<input-file-path> Path to the audio file to transcribe.
OPTIONS:
-f, --format <format> Output format ('text' or 'json'). (values: text, json; default: text)
-l, --language <language>
BCP-47 identifier for the transcription language. Defaults to the first supported system language.
--overwrite Overwrite the output file if it already exists.
-h, --help Show help information.
Look, I'm no Swift expert. Most of this code was cobbled together with the help of Gemini and ChatGPT. It works for me, but your mileage may vary.
So yeah, don't hold your breath for updates or bug fixes! 😉
This project is licensed under the MIT License.