Python API Reference

client = VoiceAgentClient(
    api_key: str,
    prompt: str = "You are a helpful AI assistant",
    voice: VoiceStyle = VoiceStyle.F1,
    language: Language = Language.ENGLISH,
    visemes: bool = False,
    tools: List[Dict] = None,
    on_transcription: Callable[[str], None] = None,
    on_response: Callable[[str], None] = None,
    on_audio: Callable[[bytes], None] = None,
    on_visemes: Callable[[list], None] = None,
    on_tool_call: Callable[[str, str, str], None] = None,  # (name, arguments, call_id)
    on_status: Callable[[str], None] = None,
    on_error: Callable[[str], None] = None,
)

client.on('transcription', lambda text: print(f"You: {text}"))

client = TTSClient(api_key: str)

client = STTClient(api_key: str, server_url: str = DEFAULT_STT_URL)

client = SpeechToTextClient(
    api_key: str,
    server_url: str = DEFAULT_STT_URL,
    language: Language = Language.ENGLISH,
    vad: str = "silero",
    on_partial_transcript: Callable[[str], None] = None,
    on_final_transcript: Callable[[str], None] = None,
    on_error: Callable[[str], None] = None,
)

@dataclass
class Viseme:
    id: int       # Character index (maps from server field `v`)
    char: str     # Normalized character (maps from server field `c`)
    timestamp: float  # Seconds from audio start (maps from server field `t`)

import os
from lokutor import simple_conversation, simple_tts, simple_transcribe, VoiceStyle

api_key = os.environ["LOKUTOR_API_KEY"]

simple_tts("Hello world", api_key, voice=VoiceStyle.F1)   # plays locally: needs lokutor[audio]
simple_conversation(api_key, prompt="You are helpful.")

with open("recording.wav", "rb") as f:
    result = simple_transcribe(f.read(), api_key)
print(result["text"])