Set up Azure Speech for Lirivo

What this setup unlocks

After Lirivo validates the Speech Resource Key and Region, Azure can be used for voice browsing, voice preview, live playback, generated-audio jobs, job refresh, and downloading finished generated audio into the local Lirivo audio library.

Lirivo does not sell Azure voice credits. Azure usage, quotas, billing, and free-tier eligibility belong to your Azure account.

Before you start

• A Microsoft account that can sign in to Azure.
• An active Azure subscription with billing set up.
• Permission to create Azure AI resources in that subscription.
• About 10 to 20 minutes if your Azure account is already ready. Account signup and payment verification can take longer.

If someone else manages Azure for you, send them this request: create a Speech-capable Azure AI resource in a region that supports neural text to speech and batch synthesis, then give me one Speech resource key and the region identifier.

Assisted setup shortcut

If you use a trusted browser agent, local coding agent, or Microsoft Copilot in Azure, you can ask it to guide you through Azure Portal. Keep the final key private: do not paste your Speech Resource Key into a public chatbot or support thread.

Prompt you can adapt:

Help me set up Microsoft Azure Speech for Lirivo.

Use Azure Portal guidance only. The goal is to create one Speech-capable
Azure AI or Foundry resource in a region that supports neural text to speech
and batch synthesis.

At the end, I need only:
1. Speech Resource Key
2. Region identifier, such as westeurope or eastus

Do not ask me to paste secrets into chat. Pause before any billing, payment,
or final create action. After the resource is created, show me where to copy
Key 1 or Key 2 and Location/Region into Lirivo.

If Azure Portal shows Microsoft Copilot, it can help explain Azure screens and generate Azure CLI commands, but you should still review any create command before it runs.

Practical setup choices

These choices are not required, but they make the Azure setup easier to understand and safer to maintain:

• Use a dedicated resource group for Lirivo Speech resources so the setup is easy to find later.
• Choose a resource name that describes the purpose and region, such as lirivo-speech-we.
• Pick a region from Microsoft's Speech region table, not only from a general Azure region list.
• Keep Key 1 and Key 2 available for rotation. If one key is exposed, regenerate that key and update Lirivo.
• Do not create a storage account, blob container, SAS URL, or Entra app registration for the first Lirivo Azure path.

The Azure pieces in plain language

Azure subscription

Human version: billing account. Microsoft term: subscription. It lets you create Azure resources and is where billing and permissions live.

Resource group

Human version: folder. Microsoft term: resource group. A name like rg-lirivo-speech is easier to recognize later.

Speech-capable Azure AI resource

Human version: cloud voice workspace. Microsoft terms may include Azure AI services, AI Foundry resource, or Speech resource.

Region

Human version: location. Microsoft term: region. Speech keys are region-bound, so the key and region must match.

Speech Resource Key

Human version: secret key for the voice workspace. Lirivo uses it to authenticate Azure Speech requests. Treat it like a password.

Batch synthesis

Human version: background audio generation. Lirivo starts a job, checks its status, then downloads the result when it succeeds.

Technical shortcut with Azure CLI

If you are comfortable with a terminal, Azure CLI can create the same setup with a few commands. This snippet is written for Azure Cloud Shell, macOS, or Linux. Windows users can run it in Azure Cloud Shell from the Azure Portal.

This creates a billable Azure resource. Review the region, resource group, name, and subscription before running the create command.

# 1. If you are using Azure CLI locally, sign in first:
# az login
#
# Azure Cloud Shell is already signed in, so skip az login there.

# 2. Confirm the selected subscription.
az account show --output table

# Optional: if you have more than one subscription, choose the right one.
# az account set --subscription "<subscription name or id>"

# 3. Choose setup names. Use a Speech-supported region.
LOCATION="westeurope"
RESOURCE_GROUP="rg-lirivo-speech"
ACCOUNT_NAME="lirivo-speech-$(date +%Y%m%d%H%M)"

# 4. Create the resource group.
az group create \
  --name "$RESOURCE_GROUP" \
  --location "$LOCATION"

# 5. Create the Speech-capable Azure AI resource.
az cognitiveservices account create \
  --name "$ACCOUNT_NAME" \
  --resource-group "$RESOURCE_GROUP" \
  --kind AIServices \
  --sku S0 \
  --location "$LOCATION" \
  --yes

# 6. Print the two values Lirivo needs.
az cognitiveservices account keys list \
  --name "$ACCOUNT_NAME" \
  --resource-group "$RESOURCE_GROUP" \
  --query "key1" \
  --output tsv

az cognitiveservices account show \
  --name "$ACCOUNT_NAME" \
  --resource-group "$RESOURCE_GROUP" \
  --query "location" \
  --output tsv

The first printed value is the key for Speech Resource Key. The second printed value is the Azure location for Region. If Azure says the account name is unavailable, choose a more unique ACCOUNT_NAME and run the create command again.

Step-by-step setup

1. Step 1

   Sign in to Azure Portal

   Open portal.azure.com and sign in. Confirm that the correct subscription is selected and active. If you do not have a subscription yet, follow Azure's account signup flow and complete any identity, phone, or payment verification it asks for.

   Azure signup screens vary by country, organization, and account type. The important result is an active subscription that can create Azure AI resources.

2. Step 2

   Create a Speech-capable Azure AI resource

   In Azure Portal, create a new Azure AI Foundry or Azure AI services resource that includes Speech support. Microsoft's current documentation points to the AI Foundry resource create flow.

   Fill in the required fields: subscription, resource group, resource name, region, and pricing tier. Choose names you will recognize later, such as rg-lirivo-speech for the resource group and lirivo-speech-we for the resource.

3. Step 3

   Choose the region carefully

   The region is not only a preference. Azure Speech keys work only with the matching region, and Lirivo needs the region to call the correct Speech endpoints.

   Use a region that supports both neural text to speech and the batch synthesis API. Examples currently listed by Microsoft include westeurope, northeurope, uksouth, eastus, eastus2, and westus2. Check Microsoft's Azure Speech regions page before choosing if you are unsure.

4. Step 4

   Review and create the resource

   Review the Azure form and create the resource. Wait for Azure to finish deployment, then open the new resource with Go to resource.

   A successful deployment record is only a record that Azure created the resource. It is not something Lirivo uses directly, and it is not a separate thing you need to renew.

5. Step 5

   Find the key and region

   Open the Speech-capable resource you created. In the resource menu, go to Resource Management and Keys and Endpoint. If the keys are hidden, choose Show Keys.

   Copy Key 1 or Key 2. Then copy the Location/Region value from the same resource. Keep both values together; a key from one region will not work with a different region.

6. Step 6

   Enter the values in Lirivo

   On iPhone, open Lirivo and go to Settings, Cloud Providers, Microsoft Azure. Paste the Azure key into Speech Resource Key, select or enter the matching Region, then save.

   Lirivo stores the key in iOS Keychain on your device and validates the key-region pair before treating Azure as configured.

7. Step 7

   Test Azure in Lirivo

   Open a text in Lirivo, choose the voice/provider control, and switch to Microsoft Azure. A working setup should let Lirivo load Azure voices, preview a voice, and use Azure for playback.

   For generated audio, use Generate Audio, choose an Azure voice, then watch the job on Jobs. When Azure finishes the job, use Download Result to bring the audio into Lirivo's local audio library.

Values Lirivo does not need

For this Azure setup, Lirivo only asks for a Speech Resource Key and Region. Do not paste these other Azure values into the Azure setup form:

Costs, quotas, and key safety

Creating the Azure resource is not the same thing as generating speech, but Azure pricing and free tiers can change. Check the current Azure pricing and quota information for your account before using cloud voices heavily.

If a Speech Resource Key is exposed, regenerate that key in Azure Portal and update Lirivo with the new key. Azure gives two keys so you can rotate one while the other is still available.

If you want to remove this setup later, clear the Azure credentials in Lirivo and delete the Azure resource or resource group in Azure Portal if you no longer need it.

Troubleshooting

Useful search terms

If you want to research the setup yourself or ask an Azure administrator for help, these terms are the important ones:

• Azure AI Foundry resource Speech
• Azure Speech Resource Key
• Azure Speech Keys and Endpoint
• Azure Speech region identifier
• Azure Speech batch synthesis API

Official references

• Create an AI Foundry resource
• Azure Speech supported regions
• Azure Speech text-to-speech REST API
• Azure batch synthesis guide
• Azure CLI documentation
• Azure CLI Cognitive Services account commands
• Azure CLI key-listing commands