The box was disabled for last-frame extract and a stuck videoBusy flag, so typing died after a finished job.
Co-authored-by: Cursor <cursoragent@cursor.com>
The right column was stuck to the viewport, so the list only had a sliver left. Both columns now stretch together and the library scrolls after that.
Co-authored-by: Cursor <cursoragent@cursor.com>
With no clip loaded, the output frame no longer reserves a full video aspect, so the library list can scroll more than one card.
Co-authored-by: Cursor <cursoragent@cursor.com>
The pre/prompt/post split renamed the variable but one console.info still used the old name, which crashed POST /api/generate on xAIGen.
Co-authored-by: Cursor <cursoragent@cursor.com>
Drop the how-to copy under duration and CFG, keep only the CFG default, and label megapixels so that wheel sits on the same row as steps.
Co-authored-by: Cursor <cursoragent@cursor.com>
The two-column library cap used overflow hidden, so the extend form was clipped with no way to reach Generate.
Co-authored-by: Cursor <cursoragent@cursor.com>
Duration now matches CFG as one wheel; tap the digits to type a length. Touch drags move at a quarter of the desktop rate.
Co-authored-by: Cursor <cursoragent@cursor.com>
Keep wrappers off the library mid-prompt so Use as input and Extend restore the same split, send stills to xAIGen as hidden inputs, and stop the two-column library from growing with the list.
Co-authored-by: Cursor <cursoragent@cursor.com>
Video and Image pickers, thumbwheels, aspect icons, and a Scale to MP paddle; extra prompts default to batch so each run starts from the original image.
Co-authored-by: Cursor <cursoragent@cursor.com>
It is the same LoRA card as the rest of the chain; the toggle only adds or removes that card so it no longer comes back after you turn it off.
Co-authored-by: Cursor <cursoragent@cursor.com>
A leftover process-remaining wake was flipping the row back to running or error, so Remove looked like it did nothing.
Co-authored-by: Cursor <cursoragent@cursor.com>
The Add picker vanished after the listing cache went empty, and switching engines wrote one stack over the other.
Co-authored-by: Cursor <cursoragent@cursor.com>
Krea image-to-image uses encode plus denoise graphs, not Klein reference latents, and still fails closed if the Krea models are missing.
Co-authored-by: Cursor <cursoragent@cursor.com>
AIGen hides the sliders, ignores snofs strengths, and strips the SNOFS node so that file cannot be queued from the public studio.
Co-authored-by: Cursor <cursoragent@cursor.com>
Generate can queue krea_v2_generate.json without sharing Klein encoders or falling back if Krea assets are missing.
Co-authored-by: Cursor <cursoragent@cursor.com>
Collapse tiles so prompts stay behind the info button. Use as input opens the matching Video or Image v2 mode with the saved sliders.
Co-authored-by: Cursor <cursoragent@cursor.com>
New sibling graph and mode so T2I does not borrow Edit, Compose, or Refine. Stills are ignored; size is width x height only.
Co-authored-by: Cursor <cursoragent@cursor.com>
