Two New Models, YouTube Import, and Voice-to-Voice

Here's what's new in Voice Creator Pro v1.6.9. Lots packed into this one.

Join the Discord

We just launched the Mortar Tribe Discord. Come hang out if you want faster support, want to help shape where VCP goes next, or just want to chat with other folks using the app.

Two New TTS Models

• NeuTTS. A lightweight English-only voice cloning model. It's a bit rough around the edges but it'll run on a potato, so it's a great option if your machine has been struggling with the bigger models.
• Kokoro. A fast multi-voice TTS model that punches way above its weight class for the speed it runs at. Probably the best choice in the app right now for narrating long books in the shortest amount of time.

The Lab (formerly Studio)

I've renamed the Studio tab to the Lab, and given it another purpose. Lab will still be the go-to place for creating voices that you can use in projects, but now I'll also drop experimental features before they're fully fleshed out to get your feedback.

First up in the Lab is voice-to-voice dictation. Speak into your mic and your words come back out re-spoken in any cloned voice you have in your library. The faster your GPU, the closer to real-time it runs. To access it, press the mic icon in the "Text to speak" box in the Clone and TTS tabs and start talking. The first run will be slower as models get loaded into VRAM but subsequent ones will be much faster. A push-to-talk shortcut can also be configured in settings.

TTS to Virtual Mic, and a Note on Accessibility

Paired with voice-to-voice is a new TTS to virtual mic integration (via VB-Cable). It routes any audio generated in VCP straight into a virtual microphone so it shows up as your mic in Discord, Telegram, Slack, OBS, or anywhere else.

This one came directly from a redditor who mentioned they were mute and were looking for a way to "talk" to their friends on Discord with a voice that sounded more like a real person than the usual robotic TTS options. That message stuck with me. This is just the first step but I want accessibility to be a real focus in VCP going forward, so if there's something that would make the app work better for you, please email me.

YouTube Import

You can now paste any YouTube URL into VCP and either transcribe the video or pull the audio straight in for voice cloning. Great for (ethically) grabbing a clip of someone speaking and turning it into a voice you can use, or for transcribing podcasts and interviews without leaving the app.

There's also a new ASR model in the lineup: Parakeet v3, which is genuinely shockingly fast and accurate.

Projects for Long-Form Audio: Major Polish Pass

The Projects feature got a lot of love this release based on feedback from people actually shipping books with it:

• M4B export with proper chapters, cover art, and metadata baked in
• Lexicon creation so you can define how names and unusual terms should be pronounced once and have them read consistently across the entire book
• Per-section controls for overriding the model, pause and gap timing, and generation parameters at the section level
• Inline section editing: add new segments and sections directly in the reader, and rename chapter titles in place

Other Improvements

• RDNA 2 GPU support. Thanks to the testing by an absolute Chad of a user (Marco), all AMD cards from RDNA 2, 3, and 4 are now fully supported.
• LAN access. The server now binds to 0.0.0.0 so you can hit VCP from your phone or another machine on the same network. You might get a Windows prompt confirming access to your network. This is safe and expected.

Pricing change in two days

Heads up: starting Thursday, May 7th, 2026 the price of Voice Creator Pro is increasing by $10. If you've been on the fence, this is your last chance to grab it at the current price. If you've already bought VCP, no worries, you won't have to pay extra.