AI Image Prompt Builder

What it does

Build structured image prompts for Midjourney, DALL-E, Stable Diffusion, Imagen, Leonardo, and any other text- to-image model. The form-based builder walks you through the components that consistently produce high-quality output:

1. Subject — what's in the image (a cat, a mountain, a person reading) plus key descriptors (specific breed, time of day, mood).
2. Style — the aesthetic (photorealistic, anime, oil painting, low-poly 3D, watercolor — pulled from the same curated list as the ai-art-style-picker).
3. Camera / framing — close-up vs wide shot, angle (low, eye-level, bird's-eye), focal length (35mm, 85mm portrait), depth of field (shallow, f/1.4).
4. Lighting — golden hour, studio strobe, volumetric, neon-lit, dramatic side-lit. Lighting often matters more than any other parameter for photographic realism.
5. Aspect ratio — 1:1 square, 16:9 landscape, 9:16 portrait/Stories, 3:2 photo, 2:3 print.
6. Negative prompt (Stable Diffusion only) — what NOT to include. "Blurry, low quality, deformed hands, bad anatomy" is the standard SD negative prompt starter.

The output is formatted for the model you pick: Midjourney uses --ar and --style flags after the prompt; DALL-E takes plain text with inline modifiers; Stable Diffusion accepts weighted terms like (detailed:1.3) and a separate negative prompt field. Each model's idiosyncrasies are baked into the formatter so you copy and paste cleanly.

How to use it

1. Fill in your subject in the first field. Be specific: 'a tabby cat sitting on a windowsill at sunset' beats 'a cat'.
2. Pick a style from the dropdown (photoreal, anime, watercolor, etc.). The list mirrors ai-art-style-picker.
3. Set camera + lighting from their dropdowns. For photorealism, these matter most. For illustration, you can leave them default.
4. Pick aspect ratio. 1:1 for Instagram, 16:9 for desktop wallpapers / video thumbnails, 9:16 for stories/reels.
5. Add negative prompt if your target is Stable Diffusion. Defaults are pre-filled for common quality issues.
6. Pick target model: Midjourney, DALL-E, SD, Leonardo, or Universal (works in all).
7. Click Copy. Paste straight into your model's prompt field. The output already includes model-specific syntax (--ar, weights, etc.).

When to use this tool

• Iterating on image generation when you don't have prompt-engineering experience yet.
• Standardizing a team's image-generation workflow with a consistent template.
• Reproducing a specific aesthetic (e.g. 'all hero images for our blog should be illustrated, golden-hour lit, 16:9') by saving a preset.
• Onboarding non-technical teammates to AI image tools without making them learn prompt syntax.

When not to use it

• Highly creative one-off prompts where the structure feels constraining — write freely.
• Models with unique syntax not covered (some niche text-to-image services have proprietary tokens) — read those models' docs.
• When the constraint isn't quality — sometimes you want a chaotic, bad-on-purpose, deliberately-weird image. The structured builder pushes toward 'good' which isn't always what you want.

Common use cases

• Quick generation during a typical workday
• Pre-decision sanity-check on inputs and outputs
• Educational use — demonstrating the underlying concept
• Onboarding a colleague who needs the same calculation/conversion