Cinematic Video Studio · compétence Open Skill
Cinematic Video Studio
When in need of realistic shorts, movies, reels, dramas, animations, fantasy, or 2D videos — from an idea, one-liner, script, or reference images; also to lip-sync a character's mouth to a voiceover or song, or to revise one stage (rewrite script, restyle a character, redo a clip, change music, remix the cut) or continue to the next episode.
Custom
5 1 1
https://d7qlwi7tl2u27.cloudfront.net/skill-covers/prod/4283939/2026-10-15/mtsarrjv.jpg
# Cinematic Video Studio A general-purpose video creation pipeline that turns an idea into a finished video across any genre (realistic shorts, movies, reels, dramas, animations, fantasy, 2D) and any format (9:16 vertical for reels/shorts, 16:9 landscape for movies/episodes). ## When to use - The user wants to create a video / short / movie / reel / drama / animation / fantasy / 2D video from an idea, one-liner, script, or reference images. - The user wants to lip-sync a character's mouth to a voiceover or song. - The user wants to revise one stage (rewrite the script, restyle a character, redo a clip, change the music, remix the cut) or continue to the next episode. ## Reference files - `references/workflow.md`: the full per-stage pipeline (script → refs → storyboard → clips → voiceover → music → final mix) with channel conventions and `name`/`entity_type` values. Read it before starting a full production. ## Steps 1. First use `write_todos` to list the stages of the production, and update the whole list as you progress. 2. Clarify with the user (ask only what's missing): genre/style (realistic / animation / fantasy / 2D), format (short/reel ~10–60s, movie, drama episode), orientation (9:16 vertical for reels/shorts, 16:9 landscape for movies), duration, dialogue language, and any reference images they want to use. 3. Read `references/workflow.md` for the detailed per-stage pipeline and channel conventions before generating. 4. Produce the script as a document (`create_document`, `entity_type='scripts'`). 5. Generate character / location / prop reference images (`generate_image`, `entity_type='characters'` / `'locations'` / `'props'`). 6. Generate storyboard frames (`generate_image`, `entity_type='storyboard_images'`). 7. Generate the video clips (`generate_video` — `ref2v` for consistent characters, `i2v` for single-image motion, `t2v` for pure text; `entity_type='storyboard_videos'`). 8. For lip-synced shots (a character's mouth moving in sync with a voiceover or song), generate with `generate_video model='ai2v'` (character image + driving audio) or `model='ref2v'` with `audio_url`; `entity_type='storyboard_videos'`. 9. Generate voiceover (`generate_tts`, `model='tts.design'`, `entity_type='voiceovers'`) and music (`generate_music`, `entity_type='music'`). 10. Assemble the final cut with `mix_clips` (`entity_type='final_videos'`), optionally burning in subtitles. 11. Summarize the result in natural language and call `set_guidance` to suggest next actions (e.g. redo a clip, change the music, remix, next episode). ## Notes - Video generation does not support NSFW/adult content on any channel — politely decline such requests. - Report failures honestly and never fabricate URLs. - Intermediate stage products (refs, storyboard, clips) will be visible on the canvas. - Remember the user's recurring preferences (genre, format, orientation, language, style) with `memory_store`; query with `memory_query` before starting so you can pick up where they left off.