ARVOX

Change your voice in real time.

Transform your voice instantly while keeping your natural accent, rhythm and pronunciation.

15 minutes free every day.

ArVox
Design preview
Model on
Turn Off Model
Microphone
Voice model
Virtual mic
VoiceSelected voice modelLocalCloud
Hear Myself

How it works

From your microphone to any app, in three steps.

  1. 01

    Speak into your microphone

    Choose any microphone connected to your computer.

  2. 02

    Select a voice model

    Pick a model from the library or one you trained yourself, then turn it on.

  3. 03

    Use it in any app

    Select the virtual microphone in the app you call, stream or record with.

Works with any application that lets you pick a microphone. Not tied to one calling or streaming program.

Voice library

Choose a voice, keep your own way of speaking.

The library lists every voice model we publish, with its language, accent, version and whether it runs locally, in the cloud or both. Filter by language, accent, category and availability.

  • Language and accent are part of every model
  • Free and premium models, clearly marked
  • New models appear as soon as we publish them
ArVox
Design preview

Voice Library

LanguageAccentCategory
Voice models appear here once published from our AI infrastructure.

Upload audio

Create a custom model

Custom voice models

Train a voice from your own recordings.

Upload clean audio of a voice you have permission to use. We create a training job on our infrastructure and show its real status until the model is ready in My Custom Voices.

  • WAV, MP3, FLAC, M4A or OGG
  • Upload progress, file size and duration before you submit
  • Status from queued to completed, reported by the training service

Upload audio

Create a custom model

I confirm that I have permission to use this voice for AI voice conversion.

Local and Cloud

Run it on your computer, or on ours.

Local Mode

Conversion runs on your own CPU or GPU. The app measures your hardware and tells you if it can keep up in real time. Included in every paid plan.

Cloud Compute

Conversion runs on our GPU servers when your computer is not fast enough. Bought separately as a daily hour allowance.

Languages and accents

Built for every language and dialect.

Languages and accents are managed in our catalog, not hardcoded in the app, so new ones can be added at any time. A language is only shown on a voice once a model for it has been trained and verified. See the current list in the voice library.

Pricing

Start free. Upgrade when you need more.

Full pricing

Free

$0forever

15 minutes of real-time conversion per day with eligible free models.

  • 15 minutes per day
  • Eligible free voice models
  • Prepare one custom voice
  • Cloud Compute available separately
Start free
Most chosen

Pro

$12per month

Unlimited Local Mode, custom voices and premium models.

  • Unlimited Local Mode
  • Premium voice models
  • Custom voice models
  • Cloud Compute available separately
Choose Pro

Business

$12per seat / month

Pro for teams with seat management. Priced per seat.

  • Unlimited Local Mode
  • Premium voice models
  • Custom voice models
  • Seat management for your team
  • Cloud Compute available separately
Set up a team

Cloud Compute

$3per daily GPU hour

A recurring daily allowance, not a one-time hour. Choose how many GPU hours you can use every day. The allowance renews each day at 00:00 UTC and unused time does not carry over.

Daily allowancePer dayAbout per 30 days
1 h / day$3$90
2 h / day$6$180
3 h / day$9$270

Get the desktop app.

One installer for Windows 10 and 11. Everything needed is included.