Fairy Princess Cooking Video · skill pública de Open Skill
Fairy Princess Cooking Video
Fairy Princess Cooking Video es una Open Skill pública en Digen. Ábrela para ver qué crea este flujo.
Series
2 0 2
https://d7qlwi7tl2u27.cloudfront.net/skill-covers/prod/3761876/2026-09-26/knn2fujm.jpg
# Fairy Princess Cooking Video Turn the fairy-princess cooking script into a polished vertical short: one consistent 18-year-old princess cooking a magical fairy meal across 16 scenes, each exactly 10 seconds, then stitched into one video with cheerful magical music. ## Reference files - `references/fairy-princess-script.md`: the full 16-scene script (location, action, dialogue, sound effects, timings). Read it before generating scenes; it is the source of truth for each scene's content. ## When to use - The user wants the full fairy-princess cooking video built from the script. - The user wants to advance or change one step only: regenerate/restyle a single scene, redo the character lock, produce the final mix, or add music/voiceover. ## Style lock (apply to every image and video prompt) Premium original 3D fairy-tale animation, ultra-HD, cute magical fantasy world, cinematic lighting, sparkling particles, beautiful princess aesthetic. Vertical portrait 9:16. ## Character lock — same princess in every scene An 18-year-old beautiful young princess: sweet youthful face, large sparkling emerald-green eyes, rosy cheeks, delicate features, natural pink lips, warm innocent smile. Long silky chestnut-brown wavy hair down her back, decorated with tiny pearl and crystal flower accessories. She wears the same elegant pastel-lavender princess dress with puff sleeves, floral embroidery, fitted bodice, flowing layered skirt, pearl belt, and delicate fairy slippers. Keep face, hairstyle, body proportions, dress, and appearance 100% consistent in every scene. ## Steps 1. First call `write_todos` to list the pipeline steps, then update it as you progress (at most one in_progress at a time). 2. **Character lock** — generate one reference image of the princess with `generate_image` (`model='t2i'`, `orientation='portrait'`, `resolution='1080P'`). Prompt = the full Character lock description above, full body, facing camera, neutral pose. Pass `name='Fairy Princess'`, `entity_type='characters'`. This image is the identity anchor for all later scenes. 3. **Scene images** — for each of the 16 scenes, read the scene's action/location from `references/fairy-princess-script.md` and generate a scene image with `generate_image` (`model='i2i'`, `orientation='portrait'`, `resolution='1080P'`), passing the character-lock image in `input_urls` so the princess stays consistent. Prompt = the scene's location + action + the Style lock, keeping the princess exactly as in the character lock. Pass `name='Scene N - <title>'`, `entity_type='storyboard_images'`. Generate all 16. 4. **Scene videos** — for each scene image, generate a 10s video with `generate_video` (`model='i2v'`, `orientation='portrait'`, `duration=10`, `resolution='720P'`), passing that scene's image in `input_urls`. Prompt = the scene's action as a short continuous motion (no dialogue text). Pass `name='Scene N video'`, `entity_type='storyboard_videos'`. 5. **Music** — generate one cheerful magical fairy background track with `generate_music` (`model='music'`), prompt describing light, sparkly, whimsical fairy-tale cooking music, ~160s to cover the full video. Pass `name='Fairy cooking music'`, `entity_type='music'`. 6. **Final mix** — stitch all 16 scene videos in order with `mix_clips`, adding the music track as audio. Pass `orientation='portrait'`, `name='Fairy Princess Cooking - Full'`, `entity_type='final_videos'`. Return the final video to the user. 7. On completion, call `set_guidance` to suggest next actions (e.g. restyle a scene, redo the character, add voiceover). ## Single-step edits If the user only wants one change, do only that step and stop: - **Regenerate/restyle one scene**: redo step 3 for that scene (and optionally step 4 for its video), keeping the character lock. - **Redo the character**: redo step 2, then redo affected scenes. - **Add voiceover**: generate per-scene dialogue with `generate_tts` (`model='tts.design'`, `voice_instruction` = sweet young female fairy voice