MastraM
Mastra•2mo ago
Rafik Belkadi

Azure Voice setup + outdated doc

Hello, i am not able to use azure voice models : 
Error: HTTP error! status: 500 - {"error":"Speech recognition failed: NoMatch - ","stack":"Error: Speech recognition failed: NoMatch - \n



here is my code: 
import { createAzure } from '@ai-sdk/azure';
import { Agent } from '@mastra/core/agent';
import { AzureVoice } from '@mastra/voice-azure';

const azure = createAzure({
  resourceName: process.env.AZURE_RESOURCE_NAME,
  apiKey: process.env.AZURE_OPENAI_API_KEY,
});



const voice = new AzureVoice({
  speechModel: {
    voiceName: 'neural', // or a specific voice like "en-US-JennyNeural"
    apiKey: process.env.AZURE_OPENAI_API_KEY,
    region: process.env.AZURE_REGION,
  },
  listeningModel: {
    voiceName: 'whisper', // or "en-US"
    apiKey: process.env.AZURE_OPENAI_API_KEY,
    region: process.env.AZURE_REGION,
  },
  speaker: 'en-US-JennyNeural', // Default voice
});

export const voiceAgent = new Agent({
  name: 'Voice Agent',
  instructions: `
  # Goal
  You are a helpful voice assistant that can help with voice related tasks.
  `,
  model: azure.languageModel('gpt-5-chat'),
  voice: voice,
});
Was this page helpful?