How long does KURACONV take to make a branded film?

Two to three weeks from brief to final cut, against the three to six months traditional production typically needs. The studio compresses the timeline by directing AI against a locked shotlist rather than generating loosely and fixing it later.

What is the difference between directing AI and prompting it?

Prompting accepts whatever a model returns. Directing sets the story first, builds a shotlist, locks character and product identity, applies continuity filters, and chooses the right tool per shot, so generation produces options that belong to one coherent film rather than disconnected clips.

Does every KURACONV project include original music?

Yes. Every project includes an original composition scored to the cut. KURACONV does not use stock music; sound design and score are part of the direction, as in São Paulo Blues, which paired an original soundtrack with its film.

From Brief to Final Cut: Inside the KURACONV Process

KURACONV moves a branded film from a single brief to a finished, scored cut in two to three weeks. Traditional production needs three to six months to cover the same ground. We compress the timeline by directing AI instead of prompting it, and by filtering every decision through a council of 22 minds before a single frame is generated. The seven steps below run on the method we call Sentimagem, built on three pillars: Presença, Engenharia, Narrativa. Here is how each stage works.

1. Discovery: the brief is the spine

Nothing starts without a brief. We do not take projects without one, and we do not make generic AI content. Discovery establishes what the film is for, who it speaks to, and the mood it must hold. Narrativa leads here: the story engine is set before any image exists. Motion follows the story; the image emerges from both. If the brief is thin, we fix the brief before we touch a tool.

2. Research: the mood is the brand

Brand Research grounds the film in something real. We study the brand's world, its references, its tone, and the texture it lives in, so a single coherent mood carries across every asset. That coherence is Presença, the first Sentimagem pillar. When the mood is right, the brand is recognizable before a logo ever appears.

3. Direction: AI is the lens, direction is the eye

This is the step most AI work skips, and it is the one that separates a film from a feed of slop. Before generation, we build the craft controls that hold a film together:

Shotlists that define every frame, angle, and movement in advance
Identity locks that keep characters, products, and faces consistent across shots
Continuity filters that track wardrobe, color, light, and props from scene to scene
Camera direction in motion, treating the model as a lens to be aimed, not a slot machine to be pulled

This is Engenharia: the camera is directed, not stumbled into. The council of 22 minds, photographers, directors, philosophers, and strategists, reviews the direction at this stage, before generation, against one standard: does this serve the story, or is it AI slop?

We direct AI, we do not prompt it. Generation is a tool. Direction is the decision.

4. Generation: the tools serve the shotlist

With direction locked, we generate against the shotlist using the right tool for each shot: GPT-Image-2, Seedance, Higgsfield, Kling, Veo. No single model is the answer; each is chosen for what a specific frame needs. Because the direction, identity locks, and continuity filters are already in place, generation returns options that belong to one film instead of a pile of unrelated clips.

5. Curation: choosing, not accepting

Generation produces candidates. Curation decides. We cut against the brief and the continuity filters, color grade for a consistent look, and discard anything that drifts in identity, light, or tone. The council standard holds through this stage too: a frame that breaks coherence is rejected even when it is beautiful on its own.

6. Original music: every project, no stock

Every KURACONV project includes an original composition. We do not use stock music. Sound design and an original score are built for the cut, not borrowed to fill it, the way São Paulo Blues paired an original soundtrack with its film. Music is part of the direction, scored to the emotional shape of the piece rather than dropped on top of it.

7. Delivery: the final cut

We deliver a finished Editorial AI Film, 15 to 89 seconds, with its color grade, sound design, and original score in place. Depending on the brief, the same pipeline produces Editorial Stills, Editorial Carousels, or a full Brand Identity. Two to three weeks from brief to final cut, on a focused slate, because we choose depth over volume.

What the process refuses

The pipeline is defined as much by what it excludes. We do not do performance ads, stock photography, projects without a brief, or generic AI content. Selected work, including Dubai Falcon and Everest, holds to the same spine: a brief first, direction over generation, an original score, and a final cut that looks intended rather than generated.

KURACONV moves from brief to final cut in two to three weeks by directing AI instead of prompting it: story first, direction before generation, a council of 22 minds filtering out slop, and an original score on every project.