Local Media → Text Transcriber (offline)

Transcribe video or audio files (anything ffmpeg can decode) to timestamped text locally (no cloud APIs).

Requirements

• macOS, Windows, or Linux
• Python 3.10+ (3.11 recommended)
• ffmpeg installed and available on your PATH

Install ffmpeg (macOS):

brew install ffmpeg

Install ffmpeg (Windows):

• winget install Gyan.FFmpeg
• or choco install ffmpeg

Install ffmpeg (Linux, Debian/Ubuntu):

sudo apt-get update
sudo apt-get install -y ffmpeg

Setup

From this repo folder:

python3 -m venv .venv
source .venv/bin/activate  # macOS/Linux
python -m pip install -U pip
pip install -r requirements.txt

On Windows (PowerShell):

py -m venv .venv
.\.venv\Scripts\Activate.ps1
python -m pip install -U pip
pip install -r requirements.txt

On Windows (cmd.exe):

py -m venv .venv
.\.venv\Scripts\activate.bat
python -m pip install -U pip
pip install -r requirements.txt

Usage

Transcribe a media file to timestamped text + JSON segments:

python -m transcriber.cli --input /path/to/video.mkv --model small --outdir ./output

Outputs:

• output/<video>.timestamps.txt
• output/<video>.segments.json

GUI (desktop app)

Launch the GUI:

python -m transcriber.gui

Notes (local-only)

• The first run will download the selected Whisper model weights into your local cache (still running on-device).
• All transcription happens locally on your machine.

Options

• --model: Whisper model name (e.g. tiny, base, small, medium, large-v3)
• --language: defaults to en
• --device: defaults to auto (tries GPU backends first, falls back to CPU)
• --compute-type: if omitted, defaults based on device (GPU: float16, CPU: int8)