---

name: nepa-tts-workspace description: Multi-engine text-to-speech workspace for the NEPA AI platform. Supports OpenAI TTS, ElevenLabs, Google Cloud TTS, and Azure Speech. Use when converting text to speech, generating voice narration, or doing voice cloning in NEPA AI.

TTS Workspace

898 lines — 5 TTS engines. Enterprise voice generation with cloning.

Prerequisites

pip install openai requests pydub google-cloud-texttospeech azure-cognitiveservices-speech boto3 numpy soundfile elevenlabs

Source

/home/billk/projects/nepa-ai-monorepo.BAK/vscode_forked/nepa-ai-backend/tts_workspace.py

Setup

cd ~/projects/nepa-ai-monorepo.BAK/vscode_forked/nepa-ai-backend
python run_nepa.py

Configure .env with your API keys (see .env.example).

Supported Engines

| Engine | Voices | Features | |--------|--------|----------| | OpenAI | alloy, echo, fable, onyx, nova, shimmer | Fast, high quality | | ElevenLabs | 1000+ voices | Voice cloning, multilingual | | Google Cloud TTS | 300+ voices | SSML, Neural2 voices | | Azure Speech | 400+ voices | Custom neural, SSML | | Amazon Polly | 60+ voices | Low latency, SSML |

Key Methods

| Method | Description | |--------|-------------| | generate_speech(text, engine, voice) | Single TTS call | | generate_from_script(script) | Multi-speaker narration | | generate_ssml(ssml, engine) | SSML-formatted speech | | clone_voice(reference_audio, text) | ElevenLabs voice cloning | | batch_narrate(texts, engine) | Batch file narration | | post_process(path, ops) | Normalize, denoise, trim silence | | convert_format(path, fmt) | MP3/WAV/OGG conversion |

Example Usage

import asyncio
from tts_workspace import TTSWorkspace

async def main():
    ws = TTSWorkspace()
    
    # OpenAI TTS - Nova voice
    result = await ws.generate_speech(
        text="Welcome to NEPA AI. The future of AI tooling starts here.",
        engine="openai",
        voice="nova"
    )
    print(f"Audio: {result['output_path']}")
    
    # ElevenLabs voice cloning
    cloned = await ws.clone_voice(
        reference_audio="my_voice_sample.mp3",
        text="This is Billy Kennedy, NEPA AI founder."
    )
    
    # Multi-speaker podcast script
    script = [
        {"speaker": "host", "text": "Welcome back to the show!"},
        {"speaker": "guest", "text": "Thanks for having me."},
    ]
    podcast = await ws.generate_from_script(script)

asyncio.run(main())

API Endpoints

POST /api/tts/generate
POST /api/tts/generate-ssml
POST /api/tts/clone-voice
POST /api/tts/batch-narrate
POST /api/tts/from-script
POST /api/tts/post-process

AXON Store

Available at: https://axon.nepa-ai.com (tts-workspace, $47) Download: /downloads/tts-workspace.zip

---

Full version: https://axon.nepa-ai.com/products