Magnific turns character sheets into moving fashion editorials
Magnific's GPT-6 Astra MCP workflow uses front, side, back, and portrait references to animate fashion looks while maintaining character identity. The resulting five-second collages link accessory close-ups back to each outfit with borders.

TL;DR
- Magnific published a GPT-6 Astra plus MCP recipe that turns four outfits into moving fashion editorials, per magnific's overview; its detailed brief specifies a five-second continuous shot magnific's five-second prompt.
- Each subject begins as a four-view reference asset, with front, side, back, and portrait views, according to magnific's character-sheet post.
- The moving layout keeps a full-body fashion look on screen while assigning live inset panels to face, fabric, accessories, and footwear, as magnific's five-second prompt lays out.
- The same character-sheet method later carried a six-member cast through a three-chapter animated short, in magnific's mini-series overview.
The Magnific platform says its MCP can run from ChatGPT, Claude, Cursor, and other agents, while OpenAI's Astra announcement lists ChatGPT Work, Codex, and the API as model surfaces. A Magnific reference-sheet guide names drifting hair and clothing as the continuity problem that multi-angle sheets are meant to solve.
MCP creative stack
The published fashion workflow names GPT-6 Astra and Magnific MCP as its tool chain. Magnific says its MCP and API consume the same plan credits as its other tools, and markets image-to-video motion control for ads, content, and film on its platform page.
Higgsfield separately showed Astra-driven animation routed from Photoshop, a parallel example of generated motion arriving inside an established visual workspace.
Four-view character sheet
Magnific's first asset is a sheet containing a full-length front view, side profile, back view, and face portrait. The examples define identity, outfit, colors, and accessories as attributes to preserve across every view.
The VERA variant explicitly uses the sheet to preserve the same person, face, outfit, and accessories in every animated panel, according to magnific's VERA prompt. That means the reference is a production asset, not just a casting image.
Five-second collage
The fashion video is a single landscape composition: one centered, head-to-shoes model stays visible as four close-ups remain on screen. magnific's five-second prompt breaks the layout into:
- A central full-body look in the middle third of the frame.
- Four live crops: face, upper-body fabric, hand or accessory, and shoes or hem.
- Rectangular inset borders and connector lines running to tracking boxes on the matching body detail.
- Synchronized micro-movements: blinking, a head turn, weight shift, accessory adjustment, and fabric moving in the breeze.
- A stable camera, with no cuts or zooms.
The graphic system changes with the outfit. One version calls for vivid orange borders and lines, and magnific's motion note says the movement stays subtle so the clothes remain the focus.
Vertical layouts
Magnific also published four vertical variations for Stories and Reels, keeping the full outfit and detail windows visible inside the narrower composition.
The original campaign was designed to adapt each look across horizontal and vertical formats, as magnific's overview puts it. The vertical examples treat reframing as a separate layout, not a crop of the landscape collage.
Sunny Six
The same reference-sheet approach appears in Magnific's Sunny Six mini-series. A character sheet and group portrait establish six stylized 3D band members before the story moves them through three chapters.
The episodes use a repeated character-continuity block, then change setting, action, camera, and audio:
- The rehearsal opens in a garage, blows an amp on the first chord, then has the band revive it. magnific's rehearsal prompt identifies Seedance 2.5 as the video tool.
- The field session moves the same cast into sunflowers for an arrival, performance, and post-show beat, per magnific's field-session prompt.
- The busking chapter starts with a pigeon audience and ends with coins landing in a flat cap, while its prompt adds performance guardrails: instrumentalists keep both hands on their instruments and nobody in the band claps. magnific's busking prompt
Magnific's follow-up says the prompts can be reused with different characters and stories magnific's follow-up.