Speech Methods

transcribe() - Speech to Text

# From file path
stt = client.transcribe("recording.wav", language="lug")
print(stt.transcript)
print(stt.latency_ms)

# From bytes
with open("audio.mp3", "rb") as f:
    stt = client.transcribe(f.read(), language="sw")

Returns an STTResult.

Supported languages: lug, ach, nyn, teo, lgg, sw, en, rw

synthesise() / synthesize() - Text to Speech

tts = client.synthesise("Oli otya?", language="lug")

# Audio URL (expires ~2 minutes)
print(tts.audio_url)

# Download immediately
import httpx
audio = httpx.get(tts.audio_url).content
with open("output.mp3", "wb") as f:
    f.write(audio)

Returns a TTSResult.

Supported languages: lug, ach, nyn, teo, lgg, sw