@mastra/voice-cloudflare

Cloudflare Voice integration for Mastra, providing Text-to-Speech (TTS) capabilities using open source speech models.

Installation

npm install @mastra/voice-cloudflare

Configuration

The module requires the following environment variables:

If using Cloudflare Native Bindings:

./wrangler.jsonc

"ai": {
    "binding": "AI",
  },

If using Cloudflare REST API:

CLOUDFLARE_AI_API_KEY=your_api_key
CLOUDFLARE_ACCOUNT_ID=your_account_id

Usage

import { CloudflareVoice } from '@mastra/voice-cloudflare';

// Native Bindings
const voice = new CloudflareVoice({
  binding: env.AI,
  listeningModel: {
    model: '@cf/openai/whisper-large-v3-turbo',
  },
});

// REST API
const voice = new CloudflareVoice({
  listeningModel: {
    apiKey: 'YOUR_API_KEY',
    model: '@cf/openai/whisper-large-v3-turbo',
    account_id: 'YOUR_ACC_ID',
  },
});

// Generate Text from an audio stream
const text = await voice.listen(audioStream);

Features

Open source models

Available Models

Speech Models

The following speech-to-text models are available:

Model	Description
`@cf/openai/whisper-tiny-en`	Lightweight English-only model
`@cf/openai/whisper`	Standard multilingual model
`@cf/openai/whisper-large-v3-turbo`	High-accuracy turbo model