Instruction-based Remix: Fiber City
A canonical prompt scaffold demonstrating native multimodal understanding through iterative instruction-based editing. This template showcases how to transform a static scene into a dynamic, physiologically synchronized environment while preserving physical consistency across all edits. The three-stage approach (base → modify → refine) ensures the model maintains causal fidelity without resorting to pixel-level substitution.
Stages
A city composed entirely of fiber-optic strands, suspended in a windless midnight atmosphere. Static establishing shot, anamorphic lens. The fibers form architectural structures — towers, bridges, and walkways — all rendered in luminous cyan and white against deep void space. No wind, no atmospheric particles, pure geometric precision.Establish the foundational visual vocabulary. The model must understand 'fiber-optic strands' as both material and structural element, not as decorative overlay.
The fibers begin to pulse like blood vessels — rhythm locked to a human breathing cadence (0.25 Hz). Pulse propagates from the city core outward in concentric waves. Each fiber maintains its structural integrity while exhibiting organic expansion/contraction. The pulse is visible as subtle brightness modulation traveling along each strand, synchronized across the entire cityscape.Critical test: the model must interpret 'pulse like blood vessels' as a temporal behavior applied to existing geometry, not as a replacement of the visual content.
Remove all background noise. Recolor to deep indigo (#0A1530). Maintain physical consistency: tension, gravity, optical bloom unchanged across edits. The pulsing behavior must continue without interruption. No new elements introduced. The color shift applies uniformly to all fiber strands while preserving their luminous quality and structural definition.The refine stage validates that the model can apply global constraints (color, noise removal) without disrupting previously established temporal behaviors (pulsing).
Compiled Prompt
role: senior_prompt_engineer @ omniveo
target_model: omni-2026.05
subject: <SUBJECT>
stage[1].base := scene(fiber_city, static_shot, geometry=architectural)
stage[2].modify := dynamic(pulse, sync=breath@0.25Hz, propagation=concentric)
stage[3].refine := constrain(noise=0, palette=#0A1530, physics=preserve)
assert: native_multimodal_understanding == true
assert: pixel_replacement == false
assert: temporal_coherence == true