Magnific publishes workflow that turns one brief into a festival campaign
Magnific published a workflow that turns a brand brief into identity assets, mockups, posters, merch, a website, event advertising, and a generated ad. A second example uses two references and Magnific's MCP with GPT-6 Astra for a smartwatch film.

TL;DR
- Magnific's magnific TRAZA demo turns one festival identity into posters, merchandise, mockups and event advertising.
- A 20-second brand film is storyboarded as five timed shots, as magnific's TRAZA prompt moves from a studio painter to a final graphic end card with an original score.
- The NOMA product spot gives its two references separate jobs: magnific's setup assigns one to character, lighting and mood, and the other to the watch's materials and details.
- The campaign output extends into a four-finish interactive landing page, which magnific's landing-page post says was built in GPT-6 Astra.
- Clip selection remains part of the process, as techhalla's walkthrough calls for generating multiple scenes and cutting the strongest into a final video.
Magnific's MCP overview lists ChatGPT alongside Claude, Cursor, Codex, VS Code, Gemini and Hermes, and exposes image, video, audio, vector and 3D generation inside those clients. Its MCP documentation says access is limited to paid plans and every MCP action consumes credits. Any multi-model chain built in Spaces can be saved as a Flow and run from a chat assistant, according to Magnific's Flows documentation.
TRAZA campaign system
Magnific's TRAZA mockup brief locks a brush-cut wordmark, ultramarine, acid lime, cream and pink, then carries its hand-drawn flower and orbit marks across:
- Venue: LED screens, a stage, entrance, kiosks and lanyards.
- Street: wheat-pasted posters.
- Merchandise: canvas totes, T-shirts and stickers.
- Campaign imagery: a studio artist, a Mediterranean garden, a portrait and a dusk-stage scene.
Five-shot festival film
The accompanying Seedance 2.5 prompt asks for a 20-second edit with no dialogue, a continuous instrumental score and five exact four-second beats:
- 0-4s: an artist paints suspended canvas in a sunlit studio.
- 4-8s: a match cut carries a drawn spiral into a garden walk.
- 8-12s: a side close-up follows the artist in late light.
- 12-16s: the camera approaches the blue sculptural stage at dusk.
- 16-20s: a hard cut lands on the TRAZA end card, held legibly for the final three seconds.
The prompt also restricts the cast, calls for natural anatomy and keeps the campaign doodles as flat 2D animation rather than glowing particles.
Magnific's ChatGPT connection post directs users to connect the service inside ChatGPT.
Two-reference watch spot
Magnific frames NOMA 01 as a cinematic watch ad and landing page made through one GPT-6 Astra plus Magnific MCP workflow.
The two inputs divide art direction cleanly. In magnific's reference setup, one image governs the character, lighting and mood, while the other governs watch materials and details. The story then moves from a corner store to a basketball court and sunrise in magnific's shot-by-shot outline.
Its 30-second template pins three variables before the scene list:
- Product: the graphite case, silver bevel, orange button and translucent burnt-orange strap must remain identical in every shot.
- Character: one 25-year-old man, wardrobe included, must remain the same throughout.
- Film language: 35mm grain, hard cuts, no narration, a single electronic score and underlying location sound.
The six cuts are convenience store, macro notification, friends outside, basketball court, rooftop dawn and a studio product hero.
Clip selection
techhalla documented a looser version of the same image-to-video loop: make a Seedream 5 Pro starting still, use it as a Seedance 2.5 reference, rewrite the prompt into additional scenes, then generate a larger clip bin for the edit.
Four-finish landing page
The NOMA demo adds a product page with four finishes, material details and an interactive watch interface. Magnific says the screen recording of that page was also made in the chat.