Svelte app to generate audiobooks locally using XTTS
-
Install Python 3.10+
-
Run start.bat (Run 2 times, first time it will install npm packages, second time it will create venv, install requirements and run the app)
- Run
start.batto start application - Upload books, speakers, RVC mdoels and indexes (RVC and index is optional)
- Press
submitbutton to upload file (Only submitted file will be uploaded) - Reload page
- Select the uploaded book and speaker from the dropdown
| Button | Description | Saves Changes |
|---|---|---|
| Narrate All | Begin generating audio for all lines in book | YES |
| Pause Narration | Pause generating audio for all lines in book | NO |
| Play All | Play all generated audios from beginning to end | NO |
| Pause Playing | Stop playing generated audios | NO |
| Undo Changes | Revert line order to last unsaved changes | NO |
| Save Changes | Saves changes made to book (Reorder\Add\Delete lines) | YES |
| Merge Book | Merge generated audios for all lines in single wav file | YES |
| Play Merged | Play merged audiobook | NO |
| Index Effect | Change the strength of the index | NO |
| Voice Pitch | Change the pitch for the generated voice | NO |
| Button | Description | Saves Changes |
|---|---|---|
| Narrate Line | Generate audio only for selected line | YES |
| Play Line | Play audio only for selected line | NO |
| Add Line Below | Add blank line below selected line | NO |
| Delete Line | Delete selected line | NO |
| Up/Down Arrows | Move selected line up/down | NO |
- Implement TTS
- Save audios in audiobooks\BOOK
- Combine audios in audiobooks\BOOK\AUDIOBOOK
- Add\Delete lines in UI (And generated audio for selected line)
- Reorder lines in UI
- Undo unsaved changes
- Save BOOK file with modified lines (Join all lines in single txt)
- Save modified lines (Delete lines in books\BOOK and split new book)
- Sync delete/modify audios and lines
- Play audio in Svelte UI
- Re-narrate selected line
- Implement RVC
- Implement multi-lang
- Might have to reload the page sometimes for changes to work
I am using coqui-ai's TTS and Jarod Mica's (Jarods Journey on youtube) rvc-tts-pipeline for the voice generation.