ai-infra

Text-to-Speech (TTS)

Convert text to natural-sounding speech with multiple providers.

Quick Start

python
from ai_infra import TTS

tts = TTS()
audio = tts.speak("Hello, world!")
audio.save("output.mp3")

Supported Providers

ProviderVoicesModelsStreaming
OpenAI6 voicestts-1, tts-1-hd✅
ElevenLabs10+ voiceseleven_multilingual_v2✅
Google5 voicesMultiple✅

Basic Usage

Auto-Detect Provider

python
from ai_infra import TTS

# Auto-detects from env vars (OpenAI → ElevenLabs → Google)
tts = TTS()
audio = tts.speak("Hello!")

Explicit Provider

python
tts = TTS(provider="elevenlabs")
audio = tts.speak("Hello from ElevenLabs!")

Voice Selection

List Available Voices

python
from ai_infra import TTS

# Get voices for current provider
voices = TTS.list_voices()
# ['alloy', 'echo', 'fable', 'onyx', 'nova', 'shimmer']

# Get voices for specific provider
voices = TTS.list_voices(provider="elevenlabs")
# ['Rachel', 'Domi', 'Bella', 'Antoni', ...]

Use Specific Voice

python
tts = TTS(provider="openai", voice="nova")
audio = tts.speak("Using the Nova voice")

# Or per-call
audio = tts.speak("Different voice", voice="echo")

OpenAI TTS

python
from ai_infra import TTS

tts = TTS(provider="openai")

# Available voices: alloy, echo, fable, onyx, nova, shimmer
audio = tts.speak("Hello!", voice="alloy")

# HD quality model
audio = tts.speak("HD quality", model="tts-1-hd")

# Standard quality (faster)
audio = tts.speak("Standard quality", model="tts-1")

ElevenLabs TTS

python
from ai_infra import TTS

tts = TTS(provider="elevenlabs")

# High-quality multilingual
audio = tts.speak("Hello!", voice="Rachel")

# Specify model
audio = tts.speak(
    "Multilingual support",
    model="eleven_multilingual_v2"
)

Google TTS

python
from ai_infra import TTS

tts = TTS(provider="google_genai")

audio = tts.speak("Hello from Google!", voice="Puck")

Audio Output

Save to File

python
audio = tts.speak("Hello!")

# Save as MP3
audio.save("output.mp3")

# Save as WAV
audio.save("output.wav", format="wav")

Get Raw Bytes

python
audio = tts.speak("Hello!")

# Get audio bytes
audio_bytes = audio.data

# Get format info
print(audio.format)  # 'mp3'
print(audio.sample_rate)  # 24000

Stream Audio

python
# Stream chunks as they're generated
for chunk in tts.stream("Long text to speak..."):
    play_audio_chunk(chunk)

Async Usage

python
import asyncio
from ai_infra import TTS

async def main():
    tts = TTS()
    audio = await tts.aspeak("Async speech!")
    audio.save("output.mp3")

asyncio.run(main())

Configuration

python
tts = TTS(
    provider="openai",
    voice="alloy",
    model="tts-1-hd",
    speed=1.0,  # 0.25 to 4.0
)

Error Handling

python
from ai_infra import TTS
from ai_infra.errors import AIInfraError, ProviderError

try:
    tts = TTS()
    audio = tts.speak("Hello!")
except ProviderError as e:
    print(f"Provider error: {e}")
except AIInfraError as e:
    print(f"TTS error: {e}")

See Also