TranscribeOne

A macOS desktop app for transcribing audio and video files with automatic speaker labels.

TranscribeOne makes it easy to turn any audio or video recording into a clean, speaker-labeled transcript. Drop in a file, click Transcribe, and get results in seconds.

Features

Transcription with Speaker Labels

Automatically detects multiple speakers and labels each line of the transcript. Works with interviews, meetings, podcasts, and any multi-speaker recording.

Speaker Identification

Enter expected speaker names before transcription and TranscribeOne will attempt to match them to voices in the recording. When automatic matching isn’t possible, use the built-in rename fields to assign names manually.

Audio & Video Support

Supports a wide range of audio formats including MP3, WAV, M4A, FLAC, OGG, WMA, and WebM. With ffmpeg installed, video files (MP4, MOV, AVI, MKV, and more) are automatically converted to audio before transcription.

Built-in Audio Player

Play back your audio directly in the app with play/pause, stop, a seek bar, and adjustable playback speed (0.5x to 2.0x).

Drag and Drop

Drag audio or video files directly from Finder onto the app window. Also supports dragging files onto the app icon in the Dock.

Auto-Save

Transcripts are automatically saved alongside the source file as <filename>-transcript.txt, and can also be copied to the clipboard or saved to a custom location.

Getting Started

Prerequisites

Install Dependencies

pip install -r requirements.txt

This installs:

Run from Source

export ASSEMBLYAI_API_KEY="your-key-here"
python transcribeone_gui.py

Build the macOS App

bash build_app.sh

This creates:

Note: The app is unsigned. On first launch, right-click the app and select Open, or go to System Settings > Privacy & Security to allow it.

Command-Line Usage

TranscribeOne also works as a CLI tool:

export ASSEMBLYAI_API_KEY="your-key-here"
./transcribeone.py recording.mp3

Output is printed to stdout with one line per utterance:

Speaker A: Welcome to the meeting.
Speaker B: Thanks for having me.
Speaker A: Let's get started with the agenda.

Pipe to a file to save:

./transcribeone.py recording.mp3 > transcript.txt

How It Works

TranscribeOne uses the AssemblyAI speech-to-text API for transcription and speaker diarization. Here’s the flow:

  1. Upload - Your audio file is uploaded to AssemblyAI’s secure servers
  2. Transcribe - AssemblyAI’s AI models convert speech to text with speaker diarization enabled
  3. Identify - If speaker names are provided, AssemblyAI’s Speech Understanding API attempts to match names to speakers based on conversational context
  4. Display - The labeled transcript appears in the app and is auto-saved to disk

About AssemblyAI

AssemblyAI provides state-of-the-art AI models for speech recognition, speaker diarization, and audio intelligence. Their API powers the core transcription engine in TranscribeOne.

Key capabilities used:

Sign up for a free AssemblyAI API key to get started.

Supported Formats

Type Formats
Audio .mp3, .wav, .m4a, .flac, .ogg, .wma, .webm
Video .mp4, .mov, .avi, .mkv, .webm, .flv, .wmv, .m4v

Video support requires ffmpeg (brew install ffmpeg).

Project Structure

transcribeone.py        # Core library and CLI tool
transcribeone_gui.py    # macOS GUI application
test_transcribeone.py   # Test suite (65 tests)
TranscribeOne.spec      # PyInstaller build configuration
build_app.sh            # Build script for .app and .dmg
requirements.txt        # Python dependencies

License

TranscribeOne is released under the CC0 1.0 Universal license (Public Domain). You are free to use, modify, and distribute it without restriction.