► Start here: https://www.skool.com/digiaistudio/classroom/2b32acf2?md=aca4a645d1ea4423a4e91a12ef0a7571 One floor plan in. A client-ready presentation out, in about an hour. Built for interior designers, architects, stagers and property agents. You leave with: - A concept board - Three room renders that match each other - A furnished plan view - A 10 second walkthrough clip One master prompt. GPT Image 2 for the stills, Seedance 2.5 for the clip. No 3D software. Every prompt is on the page. Pro members: it is in the classroom now.

https://www.skool.com/digiaistudio/classroom/cc216556?md=913a89e59427474a938ca03b7f1855c1 A poster is read from four metres. A feed ad is read in one second. An editorial spread is read for four minutes. Same design rules, six completely different jobs, and most AI output fails because nobody told the model which job it was doing. DES05 fixes that with one line. Every prompt in the course starts with DESIGN TYPE and the reading distance, and everything else follows from it: the ratio, the size of the headline as a fraction of the frame, how many words survive, where the empty space has to be, whether the finish is ink on paper or clean flat pixels. Six pages, six types, each one built twice: Page 1, THE POSTER: read from four metres. One subject, one line, silkscreen inks. Page 2, THE PACKAGE: read on the shelf, then in the hand. Three things on the face, one dominant colour, the material named. Page 3, THE FEED AD: read in one second on a phone with the sound off. One product, one message, the bottom fifth kept clear. Page 4, THE EDITORIAL SPREAD: read for minutes. Two pages, a gutter the model does not know exists, one photograph that bleeds. Page 5, THE THUMBNAIL: read at two centimetres next to eleven rivals. One object, three words, contrast you can see with your eyes half closed. Page 6, THE SLIDE: read from the back row while someone talks. One idea, one number at a third of the slide height. Page 7 is the Design Type Check, six questions before you ship any of them, and the Fix Bank for when the model ignores the job. Every page carries the base block, the advanced block, the full before and after prompt with both images, a worked example from the vaults, a swap exercise, and the why behind each rule. Pro members, it is in the classroom now. Start on page 1 and run the poster prompt once with the DESIGN TYPE line and once without. That gap is the course.

Pro members, it is in the classroom now: https://www.skool.com/digiaistudio/classroom/6becade3?md=a634b56a20ee4ddb91ba609d06f5ed51 Six codes. Thirty-six subjects. Every one shot twice. Vol.19, 20 and 21 gave you shortcut codes for information: blueprints, CFD, mission profiles. This vault turns the same idea on cinema. A cinema code is a short phrase that points GPT Image 2 at a whole archive of film craft. "Film still from" brings the anamorphic frame. "Lit only by" brings motivated single-source light. "A lone figure dwarfed by" brings scale. Type the code and you get a competent movie frame for free. Type it with the nine-block control layer and you get a frame that looks like a crew of forty made it. Every page shows both. The bare code exactly as typed, then the controlled prompt with THE SHOT, THE LIGHT, THE BLOCKING, THE WORLD, COLOUR SCIENCE, THE TEXTURE, THE RULE OF ONE and DO NOT. And under every pair, the why: the cinematography and design principle each block is built on, so you can move the codes to your own subjects instead of copying mine. Inside the classroom: Page 1, THE FILM STILL CODE: six films that do not exist, one frame each. A space western on a salt flat, a heist in a flooded library, first contact at a kopitiam at five in the morning. Page 2, THE ONE LIGHT CODE: a welder's arc, a phone screen in a blackout ward, a fridge door at three a.m. One source, nothing else. Page 3, THE COLOSSUS CODE: a dam face, a ship's bulb in dry dock, a rocket on the pad, a buried stone head the size of a hill. Scale as the emotion. Page 4, THE ANOMALY CODE: snow on one house only, a river through a metro platform, a full moon resting on the sea. Documentary physics around one lie. Page 5, THE ARCHIVE CODE: found photographs from 1899 to 1985 of things that never happened, on the right film stock for the year. Page 6, THE AFTERMATH CODE: the morning after the final, one hour after liftoff, four a.m. in the wedding hall. The story told by what is left behind.

► Start here: https://www.skool.com/digiaistudio/classroom/be46f2ca?md=be10e46f64ee433c8c1b406fe1a49298 One presenter. Thirty seconds. The product in her hand from the first frame to the last, and she talks the whole way through. That is the UGC ad every brand is paying for right now, and Vol.M is how you make six of them with one face that never changes. Vol.M is six vertical 9:16 ads, one per Vol.22 brand: a cold brew, a roaster, a running shoe, a serum, a fragrance, a granola. Each ad is one thirty-second Seedance 2.5 talking take with the full script written into the prompt, plus a ten-second product cut-in you lay over the demo beat in the edit. Same presenter in all six, built once as a bible and locked by reference. Six rules carry the whole volume: 1. The Presenter Bible. One face made once, the body made from that face, both bound to every take. Two text-to-image runs are two different people; this is how you stop that. 2. The Script Grid. Five timed beats, hook, problem, demo, proof, ask, under seventy words. Thirty seconds is fewer words than you think, and the model mumbles when you overwrite it. 3. The Hand Rule. The product is in her hand in the first frame and never leaves it. A bag that opens on camera, a pipette that drops, a cap that twists, all written as end states. 4. The Vertical Frame. 9:16 with her eyes on the upper third and the top and bottom fifths kept quiet, so captions and the endcard have room later. 5. The Cut-in. A separate ten-second product-only take, same reference, so the label is sharp when it matters. 6. The Caption Plate. No text is generated. Captions and the endcard are real type in post, on the quiet fifths you left empty. Every page carries the talking-take prompt, the cut-in prompt, the script grid, the settings, the still-frame checklist and a Fix Bank of what actually went wrong on that ad and what fixed it. Copy the prompt, swap the product, keep the presenter.

► Start here: https://www.skool.com/digiaistudio/classroom/50cda750?md=285a6f84b2e94dbb92eff470bb79ea6b ► Generate these on OpenArt: https://openart.ai/?via=digiAI (code DigiAI for 15% off plus 100 free credits) Colour and type are done. DES04 is the third leg: where things go, and why. Most AI images fail at layout, not at rendering. The product is fine, the light is fine, and the frame is still wrong: the subject sits dead centre, the text has nowhere to live, every image in a set is a different shape. This course fixes that with six systems you can write straight into a prompt: the twelve-column grid, the golden split, negative space as a designed zone, three layers of depth, scale for hierarchy, and series layouts that make six images read as one family. Every page has a Before and an After from the same product, so you see exactly what the layout instruction changed, and every prompt is on the page to copy. Built on the DES01 to DES03 rules and the Vol.22 brands, so you can run the whole set on a product you already have. Pro members: start with page 1, run the grid on any Vol.22 product, and post your Before and After below.