Quickstart

from moonshine_voice import Transcriber, ModelArch, get_model_for_language, load_wav_file

# Download and get the model
model_path, model_arch = get_model_for_language("en")

# Load audio file
audio_data, sample_rate = load_wav_file("speech.wav")

# Create transcriber
transcriber = Transcriber(
    model_path=model_path,
    model_arch=model_arch
)

# Transcribe
transcript = transcriber.transcribe_without_streaming(audio_data, sample_rate)

# Print results
for line in transcript.lines:
    print(f"[{line.start_time:.2f}s] {line.text}")

transcriber.close()

from moonshine_voice import MicTranscriber, ModelArch, TranscriptEventListener

class PrintListener(TranscriptEventListener):
    def on_line_text_changed(self, event):
        # Print updates in real-time
        print(f"\\r{event.line.text}", end="", flush=True)
    
    def on_line_completed(self, event):
        # Print final version
        print(f"\\n\u2713 {event.line.text}")

# Get model
from moonshine_voice import get_model_for_language
model_path, model_arch = get_model_for_language("en")

# Create microphone transcriber
transcriber = MicTranscriber(
    model_path=model_path,
    model_arch=model_arch
)

# Add listener
transcriber.add_listener(PrintListener())

# Start listening
print("Listening... Press Ctrl+C to stop")
try:
    transcriber.start()
    while True:
        pass  # Keep running
except KeyboardInterrupt:
    transcriber.stop()
    print("\\nStopped")

from moonshine_voice import MicTranscriber, IntentRecognizer, get_model_for_language, get_embedding_model

# Get models
asr_model_path, asr_model_arch = get_model_for_language("en")
embedding_model_path, embedding_model_arch = get_embedding_model("gemma-300m", "q4")

# Create intent recognizer
recognizer = IntentRecognizer(
    model_path=embedding_model_path,
    model_arch=embedding_model_arch,
    threshold=0.7
)

# Register commands
def on_lights_on(trigger, utterance, similarity):
    print(f"Turning lights ON (confidence: {similarity:.0%})")

def on_lights_off(trigger, utterance, similarity):
    print(f"Turning lights OFF (confidence: {similarity:.0%})")

recognizer.register_intent("turn on the lights", on_lights_on)
recognizer.register_intent("turn off the lights", on_lights_off)

# Create transcriber and attach recognizer
transcriber = MicTranscriber(
    model_path=asr_model_path,
    model_arch=asr_model_arch
)

transcriber.add_listener(recognizer)

# Start listening for commands
print("Voice command system ready")
print("Try saying: 'Please turn on the lights' or 'Switch off the lights'")
transcriber.start()