Flovvy

Changelog

New features, improvements and fixes in every Flovvy release.

🐛 Fixes

  • Voice cloning: Fixed cloning statuses displayed in a broken state.

🔧 Improvements

  • Improved TTS generation logging.

🐛 Fixes

  • Fixed rapid TTS voice toggle on and off causing errors.

✨ Features

This update includes a significant change to widget event dispatching — one of the core elements of the system — in order to implement the features listed below. Please report any issues you encounter.

  • Widget events are no longer dispatched to widgets that are sleeping (not launched but not flagged as shut down either).
  • The invoice status page now shows a clear explanation when a donation widget event is not dispatched because the streamer's widgets are not running. Previously, a generic error message was shown, which was confusing.
  • Yap! and Soundscape: A warning is now displayed at the top of donation pages when the corresponding widgets are not running, informing donors that their donation display on stream will be delayed.

🔧 Improvements

  • Restored multilingual support for voices of one of the TTS providers (specifically, used by Yap! bot Lira). Previously, this was disabled due to an undocumented API change on the provider's end.

🔧 Improvements

  • Yap!: User-generated voices and voices marked as new are now displayed at the top of the list in the donation form.
  • Soundscape: User-generated voices are now displayed at the top of the list in the donation form.
  • Yap!: Added preliminary support for the LLM model. Currently available for testing only.
  • Yap!: Adjusted LLM instructions for better avoidance of overly long responses.
  • Voice cloning: All previously cloned voices have been migrated to a new TTS model.

🐛 Fixes

  • Fixed DonationAlerts integration not reporting errors during Yap! and Soundscape partner app activation.
  • Fixed a number of issues with the voice cloning component not behaving correctly when cloning voices one after another without a page refresh.
  • Fixed cloned voices not using the new TTS model.

🔧 Improvements

  • Voice cloning: Switched to a new model for voice cloning that provides better coherence while remaining faithful to the original voice.
    You don't need to re-clone your voices, we'll automatically migrate them in the following days.

🐛 Fixes

  • Voice cloning: Fixed .mp4 files not being available for selection in the voice cloning form.

🔧 Improvements

  • Added TTS API for external uage.

🔧 Improvements

  • Voice cloning: Improved speech recognition precision, which will result in better voice cloning quality.
  • Voice cloning: Tweaked some parameters of the cloned voice to improve quality.
  • Yap!: Minor changes to bots availability and ordering logic, avatar file processing, and TTS provider support.
  • Shamelessly plugged a voice cloning banner into the main menu and the Yap! settings page.

🐛 Fixes

  • Fixed the voice cloning form not clearing after successful submission.
  • Minor fixes related to dashboard styling.

✨ Features

Ever wanted your voice to be used in donation voiceovers?
Now you can... and it's free!
  • You can clone up to 5 voices;
  • Just 15 seconds of audio is required;
  • Cloning is performed by a state-of-the-art model;
  • Voice cloning takes just a couple of minutes — after that, your voice is ready to use;
  • Cloned voices will be available in Soundscape and Yap! donations. Future releases will also bring them to standard DonationAlerts donations.
Voice cloning is already available in the Settings section of your dashboard.

🔧 Improvements

  • We're putting the finishing touches on a brand-new feature. More details coming very soon!