The voice changer and AI Voice are now free with the ready-made voices. No daily limit, no counter.
Many Steam reviews said the app wasn't truly a voice changer, and wasn't truly free. That made people understandably unhappy, and it went against the reason I made this app in the first place. Now that I have worked out my own budget, I'm proud to say both are true. I hope new players no longer feel misled when they click "Free" on Steam, and that past players give it another try.
What still needs credits
- Custom voices: every new custom voice comes with a free 7-day trial. If you like it more than the ready-made voices, any purchase of a pack or plan lets you keep it. That helps me keep the servers running without going too far into personal debt.
- The same purchase also unlocks Realtime offline, the voice changer that runs on your own graphics card, even with no internet. If you have bought anything before, both are already yours.
- Ultra realistic, the most human-sounding engine, in both the Voice Changer and AI Voice, with any voice. It still comes with free time every day, and packs and plans add more.
- Training a custom voice for Ultra realistic still uses Ultra realistic time, the same as before.
Plans
- Plans are now only for Ultra realistic time. Every plan also lets you make as many voices as you like, with up to 5 active at once instead of 1.
- Subscribed for the voice changer? You can get this month back. In the app, open Credits, then History, and press Refund next to your plan. This works until your current month ends, and you keep your custom voices and Realtime offline.
- Either way, every current subscriber also gets 2 hours of Ultra realistic as a thank-you.
Also new since the last update
- Realtime voice changing in English and Russian now drops fewer of your words. In exchange, it starts 0.2 seconds later.
- Ultra realistic sounds more natural again. Testing showed that the newer, more expressive engine ironically made voices sound much more like AI.
- New in Settings: choose the Ultra realistic voice model: V3 (default), V4 or V4 Turbo. V4 uses credits 2.75x faster than V3, and V4 Turbo only 1.5x faster.
- The app now remembers the emotion you picked last, like Shy, when you restart it.
- Training a custom voice now takes about 20 minutes in English and Russian, and about half an hour in Chinese, instead of almost 2 hours. The app shows you how long to wait.
- The free voice training is now offered after you have heard your custom voice speak.
Fixes
- Running out of Ultra realistic time now simply opens the store, without an error message.
- On networks that block part of the internet, Lines and typed messages no longer keep going silent. The app switches to its backup connection when the first one stalls.
- Training a custom voice you haven't used in a while no longer fails at the start.
- A line that is only a tone, like , now plays a matching sound on V3.
- Custom voice portraits are being made again. They stopped from October 21 to 25, and voices that missed one are getting it now.
Thank you to everyone who supported GenderShift while it had a daily limit and I was figuring things out. Your support is what made this possible.