Deletes the model for each tts option at close
Cloned voices inconsistent
App Needs Work
Response from developer
Thanks for the thoughtful feedback. You’re right that long-form cloned speech is one of the hardest parts of this kind of fully on-device voice generation, especially from very short samples. Speaklone 1.3 includes improvements for long-form and section-based generation that should help audiobook-style projects, particularly when working chapter by chapter or section by section. The app runs locally on your Mac, so quality depends on the open voice models available today. They’re powerful, but not perfect yet. As better local models become available, I plan to add them. If you try 1.3 and still run into issues, I’d genuinely appreciate hearing more at dylan@dylanmalone.com
Really awesome
Hoping to maybe see integrations, what kind? Not sure yet as I’m thinking of what this app can do given the flagship price, it could be a flagship offering.
Will update this review as I use it.
Response from developer
Thanks so much for your support! So much more to come for this app. You're exactly right that the intention is to add models to the supported list over time, though sparingly, as I'd rather put the engineering into making great features and UX and then adding a new model family from time to time when there's a substantive jump. I hope you come to feel it was a bargain over time... back to work for me! (you may already know this, but the macOS version is included - it's a universal app - so if you have a Mac, try that too, because of the increased memory on that platform it's even better).





