Make The Prompt — Free tool by HRea
Build precise AI image prompts — a clean English prompt that works everywhere, plus structured JSON for tools that support it.
Words that become images.
Three polished starting directions. One click fills the builder and generates a valid JSON plus a final English-language prompt.
AI-powered generation — describe your image and the builder fills itself. Available to HRea training participants. See our programmes at ai-hrea.com.pl
PROMPT = STARTING POINT. ITERATE, MODERATE, FIND YOUR OWN STYLE. RESPECT THE MAKERS.
Make The Prompt is a free, browser-based AI image prompt builder for designers, marketers, photographers and creators. Fill structured fields — subject, scene, lighting, camera, lens, post-processing — and instantly export a clean JSON object and a polished English text-to-image prompt ready to paste into your favourite generator.
A JSON image prompt stores every detail of your prompt — subject, composition, lighting, lens, mood, aspect ratio, negative prompt — as named fields. That makes the prompt repeatable, version-controllable and easy to tweak field-by-field. JSON prompting reduces ambiguity and gives more consistent results across tools that accept structured input such as Stable Diffusion, FLUX, GPT Image and Gemini / Nano Banana.
Try the built-in presets in the recipe gallery above: editorial product photography, cinematic portraits, abstract compositions and more. Each preset loads a complete set of builder fields you can tweak — a fast way to learn what makes a strong AI image prompt.
An AI image prompt is a text instruction that tells a text-to-image model what to generate — describing subject, composition, lighting, style and technical parameters so the model can render a precise visual.
JSON prompting structures every prompt detail (subject, lighting, lens, mood, aspect ratio) into named fields. It is repeatable, easy to edit, and reduces ambiguity, giving more consistent results across Stable Diffusion, FLUX, GPT Image and Gemini / Nano Banana.
The structured JSON and English prompt work with Gemini / Nano Banana, GPT Image, Stable Diffusion, FLUX and Google Imagen. They also work with Midjourney, which performs best with the natural-language prompt rather than JSON.
A negative prompt lists what the model should avoid — artifacts, distorted anatomy, watermarks, clutter or unwanted styles. It helps the generator focus on the desired output.
Use 1:1 for social posts, 9:16 for stories and mobile, 16:9 for widescreen and 2:3 or 4:5 for editorial portraits. Match the ratio to where the image will be used.
Yes. Make The Prompt is a free, browser-based AI image prompt builder. No payment is required.
No account, no signup, no email. Open the builder, fill the fields and copy your prompt or JSON instantly.
Fill the four core fields (subject, scene, lighting, art direction), then refine with camera, lens and post-processing. Copy the JSON for repeatable runs and the text prompt for quick generation.
Make The Prompt is built by HRea — an independent training and development company with a strong focus on modern-technology programmes. It is part of the HRea ecosystem alongside AI HRea (AI training) and copy2prompt.com (real-world prompt examples).