📝 Pre-Production and Scriptwriting

Every scene starts with a script. In Magnific Spaces, you create a Text Node with a brief scene description—the samurai on the suspension bridge.

Before any image, the director asks: what should the audience feel? Isolation, tension, and anticipation of confrontation guide angles, lighting, and composition.

Script → emotional tone → cinematic decisions. Emotion comes first.

With the AI Assistant, the script becomes a six-shot storyboard draft, showing the scene’s cinematic progression.

The storyboard defines framing, camera perspective, visual pacing, and the buildup of tension before generating detailed images.

Storyboard = direction in sketch form. Understand the scene visually before rendering.

The character is generated in a neutral studio environment to isolate the design — realistic proportions, base clothing, natural lighting, and a neutral pose.

Create a clean character reference that can later be dressed and placed in the environment, maintaining consistency across shots.

Neutral studio → design isolation → reusable reference.

Clothing and weapons are generated as separate reference objects in the studio — samurai clothing, a traditional kasa hat, a katana sword.

Rendering each object with studio lighting lets you combine them with the character later, ensuring visual consistency.

Simulated costume and prop departments in a one-person pipeline.

As in professional VFX, the character and environment are created separately. The narrow bridge above a deep canyon, with heavy rain, moving fog, and strong winds.

A narrow bridge creates vulnerability; a gigantic canyon creates cinematic scale. The environment is dramatic structure.

Atmosphere establishes scale and tension before the action.

The visual understanding of the scene becomes a cinematography plan: shot type, camera position, camera movement, and narrative purpose.

Here, AI acts as the director and cinematographer, designing how the audience experiences the scene.

Shot list = translating emotion into technical camera decisions.

With the shot list ready, we generate the final cinematic frames. The storyboard images are separated into individual shots.

Each frame becomes a still ready for AI video generation, animation, preview, and pitch.

The result is a complete pre-production package for a film scene.

The central node-based workspace, where script, character, sketch, and render are connected in a single visual workflow.

We’re not generating random images — we’re building a controlled sequence, as in real pre-production.

Nodes connect stages; Freepik is paid, but powerful for node-based pipelines.

With the Assistant, we generate rough composition sketches — focusing on framing, blocking, and pacing, not details.

The sketch works out composition and direction before investing effort in realistic rendering.

High-contrast black-and-white pencil sketch, focused on composition over micro-detail.

A second Assistant translates the sketches into realistic, colorful, rendered frames — like a previsualization step.

Separating the sketch from the render keeps the composition locked while increasing realism.

Sketch = direction; render = live-action preview.

Within the Assistant, we use references with @sketch e @character to connect the script, character, and previous results.

The references ensure that composition comes from the sketch and appearance comes from the character, without mixing their roles.

Quality depends on clear, precise prompts; iteration is normal.

Shot 1 wide (space/scale), 2 character (emotion), 3 action close-up (tension), 4 inside the car (immersion), 5 movement (energy), 6 wide again (ending).

Each shot has a defined purpose; variations in scale create rhythm, and composition guides attention.

The sequence expresses loneliness, determination, and progress.

The full arc: Space → Character → Action → Movement → Release.

A clear arc turns six disconnected shots into a scene that breathes and comes to a close.

Open at one scale, close at the same scale — closure and continuation.

A storyboard sequence of exactly 6 images, 16:9, cool, desaturated color grading — from the empty desert to the car speeding toward the horizon.

Defines the story shot by shot and the style rule every frame will inherit.

Each image generated individually; fictional, generic character.

The complete prompt for the burly 40-year-old man with a respirator mask — clothing, environment, camera, lens, lighting, color, and constraints.

A prompt structured in sections is what keeps the character consistent across every shot.

35mm lens, low angle, blue-gray color grading, generic and unidentifiable.

A single storyboard image with 6 frames in a rigid 2x3 grid, each panel a widescreen 16:9 frame in letterbox.

Critical layout rules prevent square panels and preserve the anamorphic feel of cinema.

Don’t invent shots, don’t change the order, black-and-white pencil sketch.

Recreate the storyboard with photorealistic quality using @sketch for composition and @character for appearance.

Composition, direction, and microdetail constraints prevent the model from reinterpreting or mirroring the scene.

Match angles, positioning, scale, and vanishing point—without reinterpretation.

The shot-by-shot specification: extreme wide aerial, medium frontal, over-the-shoulder close-up, interior close-up, rigid lateral tracking shot, and final wide aerial.

Shots 5 and 6 are locked — the car’s direction, vanishing point, and dust trail cannot change.

Position in the frame, angle, and negative space defined for each shot.

The final sections of the prompt: use @character exactly, follow @sketch in the environment, photorealistic style, and output of exactly 6 images.

The final rule ensures the storyboard translates directly into actual frames without changing the direction.

Same face, same mask, same coat — consistency across all 6 shots.