Generate IVR audio prompts with AI voice, per language

You don't need to record or upload audio files: write the text for each IVR prompt and Omnifox turns it into natural speech with AI, in every language you use.

Requirements

• A plan with IVR enabled (Crece or Escala).
• Text written for each language you support (Spanish, English, Portuguese) in the IVR's fields.

Steps

1. On every block that plays a message (welcome, menu, farewell, voicemail), fill in the text boxes for each language (es/en/pt) — leave a box empty only for languages you don't serve.
2. Pick a voice: on the IVR or on individual nodes you can set which ElevenLabs voice to use. The catalog of available voices is curated by an admin in the platform's AI configuration.
3. Save your changes (they stay in draft, live calls aren't affected).
4. Click Publish. Omnifox then automatically generates the final audio for each text with that voice, for all three languages, and pushes it to the phone system before the first real call arrives.
5. If you only changed the text on one block, only that audio gets regenerated — the rest is untouched.

Example

You write "Bienvenido, gracias por llamar" (es), "Welcome, thanks for calling" (en) and "Bem-vindo, obrigado por ligar" (pt) in the welcome block, choose the "Bella" voice, and publish: Omnifox generates all three audio files, ready for the next call.

Tips

• If two different IVRs use the exact same text and voice, Omnifox reuses the audio it already generated — no extra AI credit spent regenerating it.
• Keep each language's text short and natural; long messages sound better split across a few blocks.

Troubleshooting

• No audio plays for one language: that text box was left empty on that block.
• Publishing a large IVR takes a moment: several prompts are being generated in the background — wait for it to finish before test-calling.
• Changed just one word or comma: a new audio gets generated since the text changed, even if the rest stays cached.