Repository navigation
Voices can have descriptions now. Thanks @fortesoftware for the suggestion.
Downloads
| File | For |
|---|---|
breeze-tts-2-v0.1.3-windows-x64-vulkan.zip |
Windows 10 and 11, 64 bit |
breeze-tts-2-v0.1.3-linux-x64-vulkan.tar.gz |
Linux x64 with glibc 2.35 or newer (Ubuntu 22.04, Debian 12, Fedora 36 and up) |
Same as before: Vulkan builds for NVIDIA, AMD and Intel GPUs, integrated ones included, CPU fallback, AVX2 required. The Linux file is built by GitHub Actions and shows up a few minutes after the release goes live. Setup steps are in the v0.1.0 notes and the README.
What's new
- Voice descriptions and metadata. Any voice can carry an optional text description, plus extra fields of your own like
accent: britishorsource: audiobook sample. None of it changes how the voice sounds, it is there so you can tell your voices apart. - Voice library tab in the web UI. Lists every voice with its transcript. Edit the description, add or remove fields, and press SAVE CHANGES. USE picks that voice in the clone tab, and the filter box searches names, descriptions, transcripts and fields. The description also shows under the SAVED VOICE dropdown when a voice is picked, and the save row has a description box so you can write one when you save.
- Edit over HTTP.
POST /v1/voicestakesdescriptionandmetadatawhen saving,PATCH /v1/voices/<id>changes them later, andGET /v1/voices/<id>returns a single voice. Every voice listing includes both fields now. Details and examples are indocs/voices.md. - CLI.
--save-voicetakes--description, and--list-voicesprints it under each voice.
Also changed
- A voice with a description or fields is saved in a new version of the
.breezefile that v0.1.2 and older cannot read, they skip it with a warning. Voices without them are saved exactly like before, so if you need to go back to an older build, clear its description and fields first. - Saving a voice writes a temporary file first and then swaps it in, so a failed save can no longer damage an existing voice.
- The web UI tabs wrap onto two rows on phones instead of squashing.
Please report bugs in issues.