Veilscripture AI Demo · published
1. Fixed an issue that could cause the editing panels to freeze during the NPC Cast stage.
2. Fixed an issue where generation could freeze if it was stopped immediately after a message was sent.
3. Fixed an edge case involving encryption rotation.
4. Fixed a rare issue that could reset the custom LLM model folder to its default location without user confirmation.
5. Separated private Tailscale usage from Veilscripture nodes for greater control.
6. Fixed an edge case that could cause encryption to fail.
7. Fixed a shutdown error that could leave Tailscale running in the background.
8. Fixed broken CUDA runtime after previous update.
1. Settings now correctly displays the AI models currently in use.
2. Memory Bank now actively displays its current status.
3. Added more manual control over Tailscale configuration to avoid false antivirus detections.
4. Improved remote access interface.
5. Reworked Workshop Upload preview.