Azure#

Azure Speech is Microsoft's text-to-speech service. JSON2Video integrates with Azure for high-quality voiceovers, particularly when you need:

• Languages and neural voices not covered by ElevenLabs.
• Enterprise / compliance requirements that mandate Azure as the TTS provider.
• SSML support for fine-grained pronunciation, breaks, and prosody control.

How it appears in JSON2Video#

Reference Azure on a voice element by setting model to azure:

{
  "type": "voice",
  "model": "azure",
  "voice": "en-US-JennyNeural",
  "text": "Hello, this is an Azure-generated voiceover."
}

The voice field expects an Azure voice short name. See the Azure voices by language for the complete list.

Your own Azure subscription#

Using your own Azure Speech key is not available from the dashboard yet: Dashboard → Connections only accepts ElevenLabs keys today. Azure voices run on JSON2Video's Azure account, included in every plan.

SSML support#

Azure voices accept SSML input for tags like <break>, <prosody>, <phoneme>. Pass SSML via the text field.

You do not need to write the <speak> wrapper or the <voice> element — the engine adds them, using the voice from the voice field. Both of these work:

{
  "type": "voice",
  "model": "azure",
  "voice": "en-US-JennyNeural",
  "text": "A short pause <break time=\"400ms\"/> then more text."
}

{
  "type": "voice",
  "model": "azure",
  "voice": "en-US-JennyNeural",
  "text": "<speak><prosody rate=\"-6%\">Slightly slower narration.</prosody></speak>"
}

If you do supply a <speak> wrapper, its attributes are replaced with the ones Azure requires and only the markup inside it is kept. Supply a <voice> element yourself only when you want to override the voice field.

Cost#

Azure voices are included in every plan — see the credit consumption table.

See also#

• Voice element reference
• Dashboard — Connections
• Tutorial 7 — Text-to-speech voiceover
• ElevenLabs integration