try train WAN 2.1 14B LoRa again because last one was distorted and I need to make this work #ugc

think in shower to add Scenes tab to #ugc then pre-generate different scenes (which are high res images, not videos) of model talking, then they can choose one (or make new ones), add a script and that is then made into a video

make first video with #ugc

get gray dark mode in Tailwind #ugc via @marckohlbrugge

create first model training with #ugc

add create new model modal but if trained on photos #ugc

try video preview for @marckohlbrugge but it's not fast enough yet (18 seconds) #ugc

improve layout and add logo #ugc

make add model to favorites work #ugc

make new model live update while creating #ugc

make horizontal tabs in models page #ugc

fix Tailwind 4 issues in Claude Code due to CSS cache and add cachebuster in PHP #ugc

create AI influencer modal #ugc

install Tailwind CLI first time #ugc

add menus for cards #ugc

add models tab #ugc

make app dashboard for #ugc

make #ugc domain and Nginx work

provision new VPS #ugc

make Hetzner VPS for #ugc