Xport Studio vs Applio
Both run RVC voice models on your own computer, and both load the same .pth files. Applio is free, open source and gives you every dial. Xport is one installer with the rest of a vocal workflow in the same window.
The short version: Applio is the best free way to run RVC. If you are happy to handle a setup step, or you need real-time conversion, Linux or an Intel Mac, use it. Xport is for people who would rather not manage an environment: a notarized app on Mac, a normal installer on Windows, and stem splitting, key and BPM, noise removal and mastering next to the voice tools. Your models work in both, so you do not have to pick one forever.
| Xport Studio | Applio | |
|---|---|---|
| Where AI runs | Your machine | Your machine |
| Price | Free tools · Pro $79 once | Free, no paywalls |
| Source code | Closed | Open source (MIT) |
| Setup | One installer, no Python | Desktop app, or a script install with Miniconda |
| Mac | Apple Silicon, macOS 15+, notarized | Apple Silicon and Intel, not notarized yet |
| Linux | No | Yes |
| Real-time voice changing | No | Yes |
| Train your own voice | Yes · Train Clone tab | Yes, every setting exposed |
| Bring your own RVC model (.pth) | Yes | Yes |
| Music tools beyond voice | Six free tools: stems, key and BPM, noise removal, mastering, converter, trimmer | Audio separator |
Pick Applio if…
- You want RVC for free and do not mind a setup step
- You need real-time voice changing for calls or streams
- You are on Linux or an Intel Mac
- You want every training and conversion setting in front of you
- Open source matters to you
Pick Xport Studio if…
- You want to install one app and start converting, with no Python, conda or scripts
- You are on an Apple Silicon Mac and want an app macOS opens without a Terminal command
- You also need stem splitting, key and BPM, noise removal and mastering in the same app
- You want your voices kept in one organized Library
- You are fine paying $79 once for the voice tools after a 7 day trial
- You have room for a download of about 680 MB on Mac (about 400 MB on Windows) plus the 4.7 GB Voice Modeling Pack, fetched once
Feature by feature
Download the installer and open it. The Voice Modeling Pack (~4.7 GB) downloads once inside the app. There is no Python, conda or install script to run.
Two routes. The newer desktop app is a small download for Windows, Mac and Linux that sets up its own environment on first run. The classic WebUI installs from a script, which on Windows downloads Miniconda and Python 3.12 and has folder rules: no spaces in the path, not inside OneDrive, and not run as administrator. If that script stops, see our guide to Applio install errors.
Apple Silicon only, macOS 15 or later. The build is signed and notarized by Apple, so it opens like any other app.
Builds for both Apple Silicon and Intel. The desktop app's release notes say it is not Apple-notarized yet, and give a Terminal command to run if macOS reports the app as damaged.
Train Clone tab. Add 5 to 30 minutes of clean, dry vocal and press Start. The result is a .pth model on your own disk. Few settings, on purpose.
The full pipeline with each stage visible: pre-processing, extraction, training and export, plus an audio analyzer that checks a dataset before you train and TensorBoard to watch a run. More control than Xport gives you.
Drop in a vocal, pick a voice from your Library, set the pitch, convert. Built for finished vocals in a song.
Conversion with pitch tuning, cleaning, upscaling and batch processing, and a plugin system for extending it.
Not offered. Xport converts recorded audio.
Yes. Live voice changing with low-latency inference. If this is what you need, Applio is the right pick.
Voice Blender mixes two of your models into a new voice. Text to speech reads a line and converts it with your RVC model.
Both as well: a Voice Blender that merges models, and text to speech that runs through a voice model. Applio's documentation notes its Edge TTS voices need an internet connection.
Six free tools: Key/BPM Finder, Stem Splitter, Noise Remover, Mix & Master, Audio Converter, Trimmer. Free forever, processed on your machine.
An audio separator sits alongside the voice tools. Applio is focused on voice conversion, so key and BPM detection or mastering happen in other apps.
The app updates itself. Pro includes every future update.
The classic WebUI's own README says it will no longer receive frequent updates, with work focused on security patches and dependency updates. The desktop app is the active line: version 1.0.8 came out in November 2026.
Free Forever for the music tools. Voice conversion is free for 7 days, then Pro is $79 one-time. Founding is $399 one-time, capped at 100. No subscriptions.
Free, with no accounts and no paywalls. The code is MIT licensed, and the project takes donations.
You are responsible for the rights to any voice you train or convert (the in-app consent dialog records this). We don't license artist voices.
The same responsibility sits with you. Applio's Terms of Use ask users to respect copyright and privacy and to secure the rights they need.
Singers and producers who want voice conversion as one step in making a record, without maintaining a Python setup.
Tinkerers, streamers who need real-time, Linux users, and anyone who wants RVC for free with full control.
Common questions
Is Applio free?
Yes. Applio is free and open source under the MIT license, with no accounts and no paywalls. Xport Studio is free to download with six tools free forever; its voice conversion is free for 7 days, then part of Pro, $79 one time.
Is there an Applio alternative for Mac?
Xport Studio runs the same RVC .pth models on Apple Silicon Macs with macOS 15 or later. The app is signed and notarized by Apple, so it opens without a Terminal step. Applio also has Mac builds, for Apple Silicon and Intel, and is the better choice on an Intel Mac.
Why does macOS say Applio is damaged?
Applio's desktop app is not Apple-notarized yet, and macOS shows that message for apps that are not. Applio's release notes give the fix: run xattr -cr /Applications/Applio.app in Terminal and open the app again. Only do this for a copy downloaded from Applio's official GitHub releases.
Do Applio models work in Xport Studio?
Yes. A model trained in Applio is a standard RVC .pth file, and Xport loads it from disk. Models trained in Xport are .pth files too, so they move the other way as well.
Does Xport Studio do real-time voice changing?
No. Xport converts recorded vocals. If you need a live voice changer for calls or streaming, Applio has a real-time mode and is the right tool for that job.
Which sounds better?
Both run RVC, so quality depends mostly on the model and the source vocal, not the app. Start from a clean, dry vocal and match the pitch to the voice. The practical difference is how much setup and how many settings you want around the conversion.
Try Xport Studio free
Same RVC models, less setup. Load your Applio .pth files into the AI vocal converter, or train a new voice with Train Clone, both on your own machine. Voice conversion is free for 7 days, then Pro is $79 once, never a subscription.
