Hello! Thanks so much for your awesome feedback on the AI Pals foundation in the Community Hub. This demo update builds on your fantastic suggestions. There's still work to be done, but my AI Pal told me to release these improvements as soon as possible. [h3]Message Cut Off Improvements[/h3] A Continue Button will now appear if the message got cut off (Generated Tokens > Max Generated Tokens setting). And does a proper resume instead of generating a new message [img]{STEAM_CLAN_IMAGE}/45208925/858c73dd3160339a4310c337b420d45247bca996.gif[/img] [h3]Multimodal Improvements[/h3] A image-vision-enabled model has been added to the recommended models section of the library, in case you want to quickly test the vision capabilities of a smaller local model. In addition, multimodal inference should now be more reliable, and more functionality has been added to the user interface (such as taking screenshots of your screen(s)). [img]{STEAM_CLAN_IMAGE}/45208925/ef68001b7829d511293bb3d8752fe18046ba74e6.gif[/img] [h3]Inspect Context [/h3] The new Inspect Context feature now gives more insight into what the context is and its state. [h3]Guard[/h3] The red Guard message now includes category information for easier identification of potential trigger words and bug detection. [h3]Clearness Improvements[/h3] Removed some irrelevant items, renamed some wrongly named items, and moved some things for clarity + updated some descriptions. The next Demo version will revisit personality creation, provide a more detailed personality, improve the Library's generic endpoint functionality (+proper explanations on how to use it) and of course continue to add improvements based on your shared experiences and ideas!