Kult Academy is Kult Studios’ school. We walk new creatives from their first prompt to their first client, with the same method we use on brand projects.
Models change every month. What lasts is a way of looking, choosing and directing. That’s what we teach, the same way crafts have always been taught: by doing, next to people who do it for a living.
Every lesson has one goal and ends with a concrete output. By the end you have work to show, not just notes.
We teach a visual grammar that works on any model, today and tomorrow, so the next release doesn’t send you back to square one.
Shots, lenses, light, editing: the language of cinema applied to AI. It’s what separates a clip from a film.
We close on a real brief and talk about turning the skill into a service you can sell.
Start for free and go all the way to the set. Each step opens the next.
The free Kult Studios group on Skool: questions, feedback on your work, news about models and announcements of new programmes.
Free · on SkoolThree months, twelve one-hour lessons: from the platform map to an ad delivered on a real brief.
12 lessons · 3 monthsOne-to-one sessions with the founders, built around your project.
On requestThose who stand out join the studio’s productions for real clients: ads, digital characters, campaigns.
By selectionIt’s where everyone who studies with us meets: you ask, you share what you’re working on, you stay up to date on new tools. It’s also where we announce new programmes first. The community is in Italian.
× Kult Studios · official programmeFrom prompt to master.
Every lesson has one goal and one concrete output: a method to practise, built on the platform we produce with in the studio.
Platform, models, prompts and images. You learn to choose before you generate.
Higgsfield is a complete production studio.
We start from the map of the platform (Soul, Cinema Studio, Popcorn, Canvas) and the roster of video models, from Seedance to Kling and Veo. Then model routing: which engine to pick for each video based on motion, realism, length, references and native audio.
By the end you’ll haveA producer’s view of the whole suite and a cheat sheet for choosing the right model.
Your LLM becomes your production assistant.
With Higgsfield’s MCP server we connect Claude, ChatGPT and Cursor to the platform: the language model writes prompts, picks the engine, generates and manages batches of variants. Then Supercomputer, Higgsfield’s agent that plans, runs and assembles multi-step jobs. Above all, credit economy: a planned budget and no credits wasted, even when an agent is generating.
By the end you’ll haveAn AI assistant connected to Higgsfield and a budget under control.
A prompt isn’t written. It’s directed.
We take apart a high-level prompt: subject, action, setting, camera, lens, light, palette, mood. We speak the language of cinematography (shot size, focal length, aperture, camera movement, blocking), which is what models really understand. Then the iteration method: prompt layering, reference images, negative constraints, seeds and controlled variations.
By the end you’ll haveA prompt architecture you can reuse on any model, today and tomorrow.
Before the camera moves, the look is set.
Text-to-image with the right model: Soul 2.0 and Soul Cinema for fashion and film, Nano Banana Pro for references and text, Seedream for complex prompts, GPT Image for colour and typography. Inside Higgsfield Soul: presets, moodboards trained on your references and Soul HEX to lock the palette. With cinematic cameras we choose body, lens, focal length and aperture.
By the end you’ll haveA visual bible for the project: look, light and palette defined before any video.
Characters, previs, video generation and editing. The image starts to move, without losing consistency.
Your first professional asset.
Continuity is the number one problem in AI production. We solve it with Soul ID, Elements and the character sheet: turnarounds, expressions and outfits built slot by slot. Face swap and character swap put the right face on the right pose, with consent and rights cleared. Then image editing: inpaint, relight, outpaint and product placement to fix hero frames without regenerating them.
By the end you’ll haveA consistent digital cast and hero frames ready for animation.
No professional set starts without a storyboard.
We turn the script into a shot list and the shot list into a storyboard with Popcorn: 4 to 8-frame sequences that lock composition, blocking and continuity. Then we build the set in 3D Jutsu: from a prompt to a navigable 3D scene, with objects, lights and cameras animated on the timeline. Renders and animatics become references for the video models before a single credit is spent on video.
By the end you’ll haveA complete previs of your ad, from the first shot to the last.
From frame to motion: this is where the image comes alive.
Text-to-video to explore atmospheres, image-to-video with start and end frames for control. Then multiple-reference-to-video: with Seedance up to 50 references across images, video and audio; with Kling and Elements, characters and locations stay locked from one shot to the next. We work in low-resolution drafts and only finalise the best takes: credits spent on choices, not attempts.
By the end you’ll haveYour hero frames turned into clips consistent with the defined look.
Generated video isn’t final: it gets transformed and edited.
Mixed Media and presets: VFX on real footage with automatic rotoscoping, camera moves and Viral Hub effects, used with judgement. Then video editing: Kling Omni Edit to change elements with a single instruction, extending a clip forwards or backwards, Genjutsu to transfer a movement or replace an object, a person or a product. Surgical fixes instead of regenerations: fewer credits, more control.
By the end you’ll haveA library of carefully chosen presets and a method to fix any clip.
Canvas, Cinema Studio, post-production and a final brief. You work like a studio, from the first node to the delivered master.
The prompt makes the image, the image makes the video, the video makes the edit.
Canvas is Higgsfield’s node board: you connect prompts, models and references, and chain image, video and audio into a single flow, generated node by node or all at once. We save the flow as a reusable template for ad variants, character sheets and storyboards, and work on it with the team in real time.
By the end you’ll haveA node-based production template to relaunch on every new brief.
This is where you shoot: set-grade camera, lenses and lights.
We take hero frames into Cinema Studio: image-to-video, camera movesets, focal length, aperture, genre presets and lighting. We work on multi-shot coverage and the right engine for every shot, managing takes, selection and targeted retakes. Fewer generations, more control, no AI look.
By the end you’ll haveProfessional, consistent, cut-ready footage, made with a cinematographer’s method.
Footage isn’t the film. The film is born in post.
We apply a real post-production pipeline to AI footage: editing and pacing, Lipsync Studio for dialogue, voice-over and dubbing, sound design, upscaling up to 4K, colour grading and reframing for every delivery format. We use Higgsfield’s plug-ins for DaVinci Resolve, Premiere Pro and After Effects: generation, editing and upscaling right in the timeline.
By the end you’ll haveA finished master ready for delivery, with every deliverable of a commercial project.
The finish line: from the client’s brief to the delivered master.
A capstone project on a real commercial brief: treatment, look, cast, storyboard, production and finishing, using the workflow built in class. We close with a group review, like in a real agency. Then the next steps: how to keep up with new models and features, and how to turn this skill into a service you can sell.
By the end you’ll haveAn AI-native ad for your portfolio and a method that adapts to every new model.
Not a certificate in a drawer: assets you use the day after the course ends.
Made on a real brief, from treatment to master.
The visual grammar that works on any model.
Ready to relaunch on every new brief in Canvas.
So the budget goes on the final takes.
Individual sessions with the founders on creative direction, production and content systems, built around your brand or project. For creators who want to speed up and brands that want to bring production in-house.
Those who stand out join the studio’s real productions. VAULT was made this way, together with students from our 1:1 mentorship.
Watch VAULT →
No. The programme starts from the map of the platform. If you already edit you’ll move faster; if you’ve never done it, you get the method to start properly.
Yes. Access to the Kult Studios community on Skool is free: you join from the group page.
Yes. Higgsfield Pro is the name of our programme, not a platform plan: for the exercises you’ll need a Higgsfield account. In class you learn to plan credits so none go to waste.
Three months and twelve one-hour lessons, each with a concrete output. Add the time you want to spend practising: that’s where the method sticks.
We pick the people who stand out in the programmes and the mentorship for real productions with our clients. VAULT is one of those projects.
Join the free community today, or apply for the next edition of Higgsfield Pro.
