Shotcraft

How Shotcraft works

AI images give themselves away with the same tells: waxy skin, plastic surfaces, impossible lighting and lenses that never existed. Shotcraft fixes that at the prompt. You pick a shot type and a look and describe what you want; it writes a fully structured, photorealistic image prompt built the way a real photographer briefs a shoot — a specific film stock and lens, lighting with both direction and shadow, and an anti-plastic render block that kills the tell-tale AI sheen.

Real gear, not vague adjectives

Instead of “high quality, 8k, realistic,” Shotcraft names the things that actually change the render: a film stock with real colour character, a specific focal length and aperture, and a lighting setup described by where the light comes from and where it falls off.

Lighting with direction and absence

Believable images depend as much on shadow as on light. Shotcraft specifies the key light, the fill, and what stays dark — the part most AI prompts forget, and the reason so many look flat and fake.

An anti-plastic render block

Every prompt ends with a block of negative and texture cues that suppress the waxy, over-smoothed look — skin pores and fabric weave instead of airbrushed plastic.

What you can shoot

• Portraits — real skin texture and catchlights, not airbrushed mannequins.
• Product — studio-grade lighting and reflections that sell the object.
• Real estate and interiors — natural window light and correct perspective.
• Food, editorial, street, landscape and poster/graphic — each with its own gear and lighting logic.

A worked example

Ask for “a coffee brand product shot” and most tools give you a glossy, obviously-AI mug. Shotcraft turns it into a full brief: a bag of beans on a weathered oak counter, shot on an 85mm lens at f/4, warm golden-hour light raking from the left with soft shadow to the right, shallow depth of field, the colour character of a classic portrait film stock, and a render block calling for real steam, ground-coffee texture and no plastic sheen. Paste that into your image tool and the result reads as a photograph, not a render.

Frequently asked questions

Which image tools does the prompt work with?

Any modern text-to-image model — Midjourney, the major diffusion models and similar tools. The prompt is written in plain, descriptive language rather than tool-specific syntax, so you can paste it almost anywhere.

Why do AI images look fake, and how does this help?

They look fake because prompts skip the specifics real photos have: a particular lens, directional light, real texture and honest shadow. Shotcraft writes those in, and adds a render block that suppresses the plastic look most models default to.

Do I need to know photography terms?

No. You choose a shot type and a look in plain language, and Shotcraft supplies the film stock, lens, aperture and lighting for you — you can learn the terms from the prompts it writes.

Is it free?

Yes — building your prompt is free. If you hit the daily limit, paste your own Anthropic API key to keep going.