Voice - Pawa AI Docs

Voice APIs live under client.voice. See the Voice API reference.

Text to speech

Returns raw audio bytes (typically MPEG):

from pawa_ai import PawaAI

client = PawaAI()

audio = client.voice.text_to_speech.create(
    model="pawa-tts-v1-20250704",
    text="Hello, this is Pawa AI speaking!",
    voice="liora",
)

with open("output.mp3", "wb") as f:
    f.write(audio)

Supported voices include ame, liora, and ayana.

Speech to text

from pawa_ai import PawaAI

client = PawaAI()

response = client.voice.speech_to_text.create(
    files=["audio_file.wav"],
    model="pawa-stt-v1-20240701",
    language="sw",
    is_speaker_diarization=False,
)

print(response)

Optional parameters:

response = client.voice.speech_to_text.create(
    files=["meeting.wav"],
    model="pawa-stt-v1-20240701",
    language="en",
    is_speaker_diarization=True,
    prompt="Transcribe a product meeting.",
    temperature=0.0,
)

client.voice.speech_to_text.transcribe(...) is an alias for create.