Transcript

GUI Audio and video transcription using OpenAI Whisper.

Features

• Transcribe audio files (WAV, MP3, FLAC, etc.)
• Transcribe video files (MP4, MKV, AVI, etc.)
• Capture live audio from system sources
• Record from specific applications

Installation

Dependencies

Before installing, make sure you have the following dependencies:

Required:

• Python 3.10+
• GTK 4

Optional but Recommended:

FFmpeg (required for video transcription and audio capture):

# Ubuntu/Debian
sudo apt install ffmpeg

# Fedora
sudo dnf install ffmpeg

# Arch Linux
sudo pacman -S ffmpeg

# macOS
brew install ffmpeg

CUDA Toolkit (for GPU acceleration - significantly faster transcription):

# Ubuntu/Debian (CUDA 12.x)
wget https://developer.download.nvidia.com/compute/cuda/repos/ubuntu2204/x86_64/cuda-keyring_1.1-1_all.deb
sudo dpkg -i cuda-keyring_1.1-1_all.deb
sudo apt update
sudo apt install cuda-toolkit-12-1

# Fedora
sudo dnf install cuda

# Arch Linux
sudo pacman -S cuda

# Verify installation
nvcc --version
nvidia-smi

Note: CUDA requires an NVIDIA GPU. For AMD/Intel GPUs, the app will automatically fall back to CPU mode.

From Source

python3 -m venv env
source env/bin/activate
pip install -r requirements.txt
python3 main.py

Flatpak build

flatpak-builder --user --install build com.conv.transcript.json
flatpak run com.conv.transcript

Usage

GUI: Run python3 main.py

CLI:

# Transcribe a file
./transcription/transcript.py audio.mp3 -m base -l en -o output.txt

# Capture from audio source
./transcription/transcript.py --capture --device "source-name"

# Capture from application
./transcription/transcript.py --capture-app

Feature Requirements

Feature	Requires
Audio file transcription	Python + Whisper (included)
Video file transcription	FFmpeg
Audio capture/recording	FFmpeg + PulseAudio/PipeWire
GPU acceleration	NVIDIA GPU + CUDA Toolkit

Without FFmpeg: Audio files only (no video or capture)
Without CUDA: CPU mode only (slower but works on any system)