Imagine Art 1.5 Pro Prompt Guide

Prompts That Produce Professional Results

Unlike general-purpose image generators that favor stylistic interpretation, ImagineArt 1.5 Pro Preview optimizes for realism. Skin textures render with subsurface scattering characteristics. Fabric displays appropriate drape and weave patterns. Metal surfaces exhibit correct specular highlights and reflection properties. These capabilities emerge from prompts that communicate in technical rather than abstract terms, treating the model as a virtual photographer who requires specific direction rather than loose inspiration.

Text-to-image models interpret prompts through language encoders that process information sequentially, prioritizing elements based on their position and specificity.¹ ImagineArt 1.5 Pro Preview exploits this architecture to generate 4K images with photographic fidelity, accurate typography, and intricate surface textures. The model responds particularly well to prompts structured around professional photography conventions, where lighting direction, focal length, and material descriptions follow established visual language.

API Implementation

The model accepts three parameters through the fal API: prompt text, aspect ratio selection, and an optional seed value for reproducible results.

import fal_client

result = fal_client.subscribe(
    "imagineart/imagineart-1.5-pro-preview/text-to-image",
    arguments={
        "prompt": "A photorealistic portrait...",
        "aspect_ratio": "2:3",
        "seed": 42
    }
)

image_url = result["images"][0]["url"]

The response returns an images array containing generated outputs. Each image object includes url, width, height, and content_type fields. The aspect ratio parameter shapes how the model composes visual elements within the frame, with options including 1:1, 16:9, 9:16, 4:3, 3:4, 3:1, 1:3, 3:2, and 2:3. The seed parameter enables reproduction of outputs when iterating on prompt variations, though results may vary slightly across API versions.

fal^{MODEL APIs}

The fastest, cheapest and most reliable way to run genAI models. 1 API, 100s of models

Build

fal^SERVERLESS

Scale custom models and apps to thousands of GPUs instantly

Deploy

fal^COMPUTE

A fully controlled GPU cloud for enterprise AI training + research

Train

Hierarchical Prompts

Effective prompts follow a prioritized sequence: subject definition, compositional framing, lighting specification, and technical parameters.² Models weight early prompt elements more heavily during initial denoising stages, then shift toward image quality refinement in later stages. This processing pattern means opening with your most critical subject details produces better alignment than burying them mid-prompt.

A functional baseline prompt: "A young woman with wavy brown hair, natural lighting, soft focus background." This establishes subject, light source, and depth relationship in three clear statements. The model receives unambiguous direction without conflicting interpretations.

For commercial applications requiring precise control, prompts expand across seven dimensions:

Subject description: age indicators, physical features, expression, wardrobe
Compositional framing: camera angle, perspective type, rule of thirds positioning
Lighting characteristics: direction, quality (hard or soft), color temperature
Material properties: fabric type, surface finish, texture descriptions
Color palette: dominant tones, contrast levels, mood associations
Camera specifications: focal length, aperture, depth of field behavior
Environmental context: setting details, background elements, atmospheric conditions

Professional Portrait Example

Portrait photography prompts benefit from terminology borrowed directly from studio practice:

"A photorealistic close-up portrait of a woman in her late twenties with shoulder-length auburn hair and hazel eyes. Sharp focus on facial features with f/2.8 depth of field. Soft natural window light from camera left creating subtle shadows. She wears a minimalist silver necklace. Background shows a softly blurred modern interior with warm beige tones. Shot with 85mm lens equivalent, capturing fine skin texture and individual hair strands."

Each sentence addresses a specific requirement. The opener defines subject and age range. The second sentence establishes technical camera behavior. Lighting direction and quality follow. An accessory detail grounds the portrait in physical specificity. Background treatment and color temperature complete the environmental context. The closing sentence reinforces technical expectations while flagging the detail level required.

This structured approach prevents the model from making interpretive decisions that conflict with creative intent. When prompts leave ambiguity, the model defaults to its training distribution, which may not match your target aesthetic.

Aspect Ratio Selection

Frame dimensions influence compositional choices before the model generates a single pixel. Vertical ratios allocate space for headroom in portraits. Horizontal ratios accommodate scene-setting elements in environmental shots.

Aspect Ratio	Primary Applications	Compositional Behavior
2:3, 9:16	Portrait photography, mobile content	Vertical framing with natural headroom
16:9, 3:2	Landscape photography, cinematic frames	Horizontal emphasis for editorial layouts
1:1	Product shots, social media	Centered subjects with balanced negative space
3:1, 1:3	Panoramic landscapes, vertical banners	Extended compositions for architectural subjects
4:3, 3:4	Traditional print formats	Classical proportions for standard frames

Selecting the appropriate ratio before writing your prompt avoids compositional conflicts. A portrait prompt paired with 16:9 forces the model to reconcile competing spatial requirements, often producing awkward cropping or inappropriate background extension.

Material and Surface Specification

Photorealistic output depends on accurate material descriptions. Generic terms like "shiny" or "soft" provide insufficient direction. Instead, specify actual material properties:

Metallic surfaces: "brushed aluminum with matte finish," "polished chrome reflecting ambient light"
Fabric textures: "chunky cable-knit wool," "lightweight linen with natural creasing"
Organic materials: "walnut burl with figured grain pattern," "marble with grey veining"
Skin rendering: "natural skin texture with visible pores," "sun-warmed complexion with subtle freckling"

These specifications activate the model's photorealistic rendering pathways, producing surfaces that exhibit physically plausible light interaction rather than stylized approximations.

Lighting Direction

Photography lighting terminology translates directly into prompt language. Specify light source position relative to camera: "key light from camera left," "backlight creating rim illumination," "overhead softbox producing even coverage." Color temperature descriptions refine mood: "warm golden hour light," "cool overcast daylight," "tungsten interior lighting."

Combining position and quality yields precise control:

Portrait lighting: "Rembrandt lighting with soft fill from camera right"
Product photography: "gradient background with overhead strip softbox"
Environmental shots: "dappled forest light through canopy"
Dramatic portraits: "single hard light from above, deep shadows below cheekbones"

Lighting prompts work best when paired with complementary shadow descriptions. The model interprets "soft shadows" differently from "hard-edged shadows," producing distinct tonal gradations that affect overall image mood.

Iterative Refinement

The seed parameter enables systematic iteration: generate an initial output, identify specific deficiencies, adjust prompt language targeting those issues, then regenerate with the same seed to isolate the effect of your changes. Note that identical seeds produce consistent but not guaranteed identical results across different API versions or infrastructure updates.

Common refinement patterns include adding specificity when outputs feel generic, removing conflicting descriptors when elements compete for visual attention, and reordering prompt segments when priority elements receive insufficient emphasis.

Typography Rendering

ImagineArt 1.5 Pro Preview handles text elements with greater accuracy than many comparable models. For prompts requiring typography, specify exact content within quotation marks: "neon sign reading 'OPEN' in red letters," "book spine showing title 'The Great Novel' in serif typography." Best results occur with short text strings of three to four words. Longer passages may exhibit degraded legibility.

Production Considerations

At $0.045 per image, ImagineArt 1.5 Pro Preview is a premium option for commercial workflows requiring high-fidelity 4K output. Generation times vary with queue depth. For batch processing or high-volume applications, consider implementing queue-based submission patterns rather than synchronous requests. Consult the fal documentation for webhook integration and asynchronous workflow patterns.