Pashto STT (local)

Minimal refactor of your single-file server into a small package.

Quick start

1. Create and activate a virtual environment (Windows PowerShell):

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

2. Run the server (binds to all interfaces by default):

python WebServer.py

Files added

• pashto_stt/asr.py — ASR model wrapper
• pashto_stt/audio.py — small helper to parse Gradio audio tuples
• pashto_stt/ui.py — builds the Gradio interface
• app.py — application entrypoint
• WebServer.py — lightweight compatibility wrapper

Background-music removal

pashto_stt.audio now offers remove_background_music(data, sample_rate, method="auto").

• Method auto will try the demucs CLI (recommended) if installed and fall back to noisereduce spectral gating otherwise.
• To use Demucs install it separately (it is optional and may require ffmpeg):

pip install demucs
# and ensure ffmpeg is installed on PATH

If you don't install Demucs, the repo will use noisereduce for a lightweight noise-reduction fallback.