voicepad 0.2.1

Command-line interface for voice recording and GPU-accelerated transcription.

voicepad

Terminal interface for VoicePad — record audio and transcribe it locally with a single keypress.

Install

# Run without installing
uvx voicepad

# Install permanently
uv tool install voicepad

Usage

voicepad

That's it. On first run, the interactive TUI opens an onboarding flow so you can choose a microphone and a starter Whisper model before the first download begins.

Key	Action
Space	Start / stop recording
c	Copy transcription to clipboard
Tab	Switch tabs (Record / History / Settings)
q	Quit

Configuration

Settings are stored at ~/.config/voicepad/voicepad.yaml and can be changed from the Settings tab inside the app. Some advanced options remain config-only.

transcription_model: turbo
transcription_device: auto
transcription_compute_type: auto
input_device_index: null
recordings_path: ~/.config/voicepad/data/recordings
markdown_path: ~/.config/voicepad/data/markdown

# Streaming / chunking
min_chunk_s: 15.0
max_chunk_s: 29.0
overlap_s: 0.5
silence_threshold_ms: 1000
min_fresh_speech_duration_s: 0.25

# Inference cleanup / prompting
initial_prompt: "Hello. This is a transcription with proper punctuation, capitalization, and grammar."
text_postprocessing_enabled: false
no_speech_threshold: 0.6
hallucination_max_repetitions: 3

# VAD tuning
vad_threshold: 0.5
vad_min_speech_duration_ms: 250
vad_speech_pad_ms: 30

The UI intentionally exposes a short curated model list for most users. If you want advanced model IDs or lower-level transcription tuning, edit voicepad.yaml directly.

Documentation

voicepad.hyperoot.dev