794d923bf9
Replace the fixed 10s post-wake-word listen window with a VAD-driven loop. Once at least 500ms of speech have been seen, 1.2s of silence closes the window. Hard cap at 15s as a safety net. Tunable in config.py via VAD_AGGRESSIVENESS, COMMAND_END_SILENCE_MS, COMMAND_MIN_SPEECH_MS, COMMAND_MIN_LISTEN_MS, COMMAND_MAX_LISTEN_MS.
28 lines
298 B
Text
28 lines
298 B
Text
# mic capture
|
|
pvrecorder
|
|
|
|
# Vosk
|
|
vosk
|
|
|
|
# Voice activity detection (smart command-end)
|
|
webrtcvad-wheels
|
|
|
|
# Groq / OpenAI-compatible LLM
|
|
openai>=1.0
|
|
|
|
# Silero stuff
|
|
torch
|
|
torchaudio
|
|
omegaconf
|
|
|
|
# etc
|
|
numpy
|
|
simpleaudio
|
|
rich
|
|
fuzzywuzzy
|
|
python-Levenshtein
|
|
comtypes
|
|
pycaw
|
|
sounddevice
|
|
PyYAML
|
|
python-dotenv
|