Skip to content

Overview

Getting Started

Welcome to ARTVoice Accelerator

Build real-time voice agents on Azureโ€”from your first deployment to production.


Choose Your Path

flowchart TD START([New to the project?]) --> Q1{Have Azure resources?} Q1 -->|No| QUICK[๐Ÿ“ฆ Quickstart<br/>Deploy everything with azd up] Q1 -->|Yes| LOCAL[๐Ÿ’ป Local Development<br/>Run backend/frontend locally] QUICK --> DEMO[๐ŸŽฎ Demo Guide<br/>Try the voice agent] LOCAL --> DEMO DEMO --> NEXT{What's next?} NEXT --> ARCH[๐Ÿ“ Architecture<br/>Understand the system] NEXT --> CUSTOM[๐Ÿ”ง Customize<br/>Build your own agents] style START fill:#e8f5e9 style QUICK fill:#e3f2fd style LOCAL fill:#fff3e0 style DEMO fill:#f3e5f5

Step-by-Step Guides

Step Guide Time Description
0๏ธโƒฃ Prerequisites 5 min Install required tools (one-time setup)
1๏ธโƒฃ Quickstart 15 min Deploy to Azure with azd up
2๏ธโƒฃ Local Development 10 min Run backend/frontend on your machine
3๏ธโƒฃ Demo Guide 10 min Create profiles, test agents, explore features

Quick Answers

How long does deployment take?

~15 minutes for complete infrastructure + application deployment via azd up.

Do I need a phone number?

No โ€” browser-based voice works without a phone number.

Phone numbers are only needed for PSTN (telephone) integration. See Phone Number Setup.

What Azure resources are created?
Category Services
AI Azure OpenAI, Speech Services, VoiceLive
Communication Azure Communication Services
Data Cosmos DB, Redis, Blob Storage
Compute Container Apps, App Configuration
Monitoring Application Insights, Log Analytics
Which Azure regions are supported?

The accelerator requires several Azure services to be available in the same region. Here's a breakdown:

For the best experience, deploy to regions with full service coverage:

Region OpenAI Speech ACS Redis Cosmos DB
East US โœ… โœ… โœ… โœ… โœ…
East US 2 โœ… โœ… โœ… โœ… โœ…
West US 2 โœ… โœ… โœ… โœ… โœ…
West US 3 โœ… โœ… โœ… โœ… โœ…
Sweden Central โœ… โœ… โœ… โœ… โœ…
UK South โœ… โœ… โœ… โœ… โœ…

Required models: gpt-4o, gpt-4o-realtime-preview

Model Regions
gpt-4o East US, East US 2, West US, West US 3, Sweden Central, UK South, France Central, Japan East, Australia East
gpt-4o-realtime East US 2, Sweden Central, West US 2

Azure OpenAI Model Availability

Required for STT/TTS in Custom Cascade mode.

Available in 40+ regions including all recommended regions above.

Speech Services Regions

VoiceLive SDK (Preview)

Currently limited to: East US 2, Sweden Central, West US 2

VoiceLive Preview Regions

ACS is a global service (not region-specific), but resources are created in a data location.

Data Location Regions
United States East US, West US 2
Europe UK South, France Central, Sweden Central
Asia Pacific Australia East, Japan East, Southeast Asia

ACS Data Residency

Available in 30+ regions. All recommended regions above are supported.

Azure Cache for Redis Regions

Cosmos DB is available in all Azure regions with multi-region write support.

Cosmos DB Regions

Region Selection Strategy

  1. Start with East US 2 or Sweden Central for full feature support including VoiceLive
  2. Use paired regions for disaster recovery (East US โ†” West US)
  3. Check quotas in your subscription before deploying
Can I use existing Azure resources?

Yes! Skip azd up and configure .env manually. See Local Development - Legacy Setup.


Getting Help

Resource When to Use
Troubleshooting Common issues and solutions
GitHub Issues Bug reports
GitHub Discussions Questions and community help

Continue Learning

After getting started, explore these areas:

Topic Guide
Understand the architecture Architecture Overview
Customize agents Agent Framework
Production deployment Deployment Guide
Add phone support Phone Number Setup