Built for controlled, film-set precision. You specify genre, color palette, lighting style, and real camera specs (body, lens, focal length, aperture, moveset) exactly like briefing a DOP. Best when the client needs a specific, repeatable cinematic look.

Works on a 5-7 element structure: subject, action, camera motion, environment, style, and optional dialogue. Supports up to 6 shots in one generation with character-anchoring across shots, so the same character stays consistent scene to scene. Dialogue must follow [Character, voice]: "line" so Kling's audio model reads tone correctly.

Rewards a clear goal over rigid parameters. Five modes cover different jobs: Reference-Based (match a look), Continuation (extend a shot), Expand Shot (widen the frame), Edit Shot (change an element), Transformation (morph one thing into another). Describe intent in plain language rather than filling technical fields.

SCALS stands for Subject, Composition, Action, Location, Style. Each element gets a short phrase, and the combined prompt should stay under roughly 75 words. Brevity keeps the model focused instead of diluted.

Supercomputer is not template-fillable. It is an agent that reads a natural-language brief and decides on its own which of the above tools (or combination of them) to run, in what order, with what parameters. Because it has no exposed API, we compose the brief here and paste it into Higgsfield's own Supercomputer chat.