A fast, accurate, on-device AI library for building interactive voice applications. Join our Discord to get help and support.
pip install moonshine-voice# Listens to the microphone, logging to the console when there are
# speech updates.
python -m moonshine_voice.mic_transcriber
"""Transcribes live audio from the default microphone"""
import time
from moonshine_voice import (
MicTranscriber,
TranscriptEventListener,
get_model_for_language,
)
# This will download the model files and cache them.
model_path, model_arch = get_model_for_language("en")
# MicTranscriber handles connecting to the microphone, capturing
# the audio data, detecting voice activity, breaking the speech
# up into segments, transcribing the speech, and sending events
# as the results are updated over time.
mic_transcriber = MicTranscriber(
model_path=model_path, model_arch=model_arch)
# We use an event-driven interface to respond in real time
# as speech is detected.
class TestListener(TranscriptEventListener):
def on_line_started(self, event):
print(f"Line started: {event.line.text}")
def on_line_text_changed(self, event):
print(f"Line text changed: {event.line.text}")
def on_line_completed(self, event):
print(f"Line completed: {event.line.text}")
listener = TestListener()
mic_transcriber.add_listener(listener)
mic_transcriber.start()
print("Listening to the microphone, press Ctrl+C to stop...")
while True:
time.sleep(0.1)If you have a different source you're capturing audio from you can supply it directly to a transcriber.
"""Transcribes live audio from an arbitrary audio source."""
from moonshine_voice import (
Transcriber,
TranscriptEventListener,
get_model_for_language,
load_wav_file,
get_assets_path,
)
import os
from typing import Iterator, Tuple
def audio_chunk_generator(
wav_file_path: str, chunk_duration: float = 0.1
) -> Iterator[Tuple[list, int]]:
"""
Example function that loads a WAV file and yields audio chunks.
This demonstrates how you can integrate your own proprietary
audio data capture sources. Replace this function with your own
implementation that yields (audio_chunk, sample_rate) tuples.
Args:
wav_file_path: Path to the WAV file to load
chunk_duration: Duration of each chunk in seconds
Yields:
Tuple of (audio_chunk, sample_rate) where:
- audio_chunk: List of float audio samples
- sample_rate: Sample rate in Hz
"""
audio_data, sample_rate = load_wav_file(wav_file_path)
chunk_size = int(chunk_duration * sample_rate)
for i in range(0, len(audio_data), chunk_size):
chunk = audio_data[i: i + chunk_size]
yield (chunk, sample_rate)
model_path, model_arch = get_model_for_language("en")
transcriber = Transcriber(
model_path=model_path, model_arch=model_arch)
stream = transcriber.create_stream(update_interval=0.5)
stream.start()
class TestListener(TranscriptEventListener):
def on_line_started(self, event):
print(f"{event.line.start_time:.2f}s: Line started: {event.line.text}")
def on_line_text_changed(self, event):
print(
f"{event.line.start_time:.2f}s: Line text changed: {event.line.text}")
def on_line_completed(self, event):
print(f"{event.line.start_time:.2f}s: Line completed: {event.line.text}")
listener = TestListener()
stream.add_listener(listener)
# Feed audio chunks from the generator into the stream.
wav_file_path = os.path.join(get_assets_path(), "two_cities.wav")
for chunk, sample_rate in audio_chunk_generator(wav_file_path):
stream.add_audio(chunk, sample_rate)
stream.stop()
stream.close()We also provide voice command recognition using the IntentRecognizer module. It captures transcribed audio from a MicTranscriber and invokes callback functions that match your programmed intents. Since it relies on an embedding model, you can use a helper function to get started:
from moonshine_voice import (
MicTranscriber,
IntentRecognizer,
ModelArch,
EmbeddingModelArch,
get_embedding_model,
get_model_for_language
)
# Download and load the embedding model for intent recognition
embedding_model_path, embedding_model_arch = get_embedding_model()Next, create a recognizer and register your intent callbacks:
intent_recognizer = IntentRecognizer(
model_path=embedding_model_path,
model_arch=embedding_model_arch
)
def on_lights_on(trigger: str, utterance: str, similarity: float):
"""Handler for turning lights on."""
print(f"\n💡 LIGHTS ON! (matched '{trigger}' with {similarity:.0%} confidence)")
def on_lights_off(trigger: str, utterance: str, similarity: float):
"""Handler for turning lights off."""
print(f"\n🌑 LIGHTS OFF! (matched '{trigger}' with {similarity:.0%} confidence)")
intent_recognizer.register_intent("turn on the lights", on_lights_on)
intent_recognizer.register_intent("turn off the lights", on_lights_off)Finally, create a MicTranscriber, connect it to your IntentRecognizer, and start the audio stream:
# Get the transcription model and initialize a MicTranscriber
model_path, model_arch = get_model_for_language("en")
mic_transcriber = MicTranscriber(model_path=model_path, model_arch=model_arch)
# The intent recognizer will process completed transcript lines and invoke trigger handlers
mic_transcriber.add_listener(intent_recognizer)
mic_transcriber.start()
try:
while True:
time.sleep(0.1)
except KeyboardInterrupt:
print("\n\nStopping...", file=sys.stderr)
finally:
intent_recognizer.close()
mic_transcriber.stop()
mic_transcriber.close()The framework currently supports English, Spanish, Mandarin, Japanese, Korean, Vietnamese, Arabic, and Ukrainian. We are working on wider language support, and you can see which are supported in your version by calling supported_languages(). To use a language, request it using get_model_for_language() passing in the two-letter language code. For example get_model_for_language("es") will download the Spanish models and pass the information you need to create Transcriber objects using them.
For more information, see the main Moonshine Voice documentation.
The code and English-language models are released under the MIT License - see the main project repository for details. The models used for other languages are released under the Moonshine Community License.