Pal Engine - Desktop Companion · beta-test, public · first seen
Voice add-ons and Health Check
- Voices sorted by language. Female, Male and Neutral voice lists show only that kind, grouped by language and region, for example Bulgarian (Bulgaria) or English (United Kingdom). The character Voice page is grouped the same way.
- Add-ons update themselves. When Edge TTS or TikTok TTS has a new version, the notification bell tells you. Click it, then Update in Settings › Integrations. Your default voice and each character's voice stay as they were.
- Health Check has two tabs. Dependencies can now remove models you no longer use (Remove or Clean unused) and shows how much space that frees. Memory shows how much RAM and graphics memory Pal Engine uses, per process, and what each helper has loaded.
- Cancel downloads. Component downloads have a Cancel button in the Downloads list.
- Your audio devices apply right away. Replies always play on the output device you chose, also in Mini-Chat, and a new microphone works at once.
- Voice Lock in Voice AI Settings. It's now there as well as in Settings › Voice, and the two stay in sync. While Pal Engine learns your voice, the character stops listening.
- "Neutral" instead of "Non-binary" for character gender, and for the third voice slot.
- Fixed: clicking Download twice on a speech model could end in a "Failed" download.
- Fixed: with the microphone on, the first greeting no longer waits for the speech model to load.
| Depot | New manifest | Previous manifest |
|---|---|---|
| 3868881 | 6773312324076247933 | 881060927180605109 |