ForkFree

Prompts / Images and video

Turn Rough Ideas Into JSON Image Briefs

You get a clean, fully specified JSON brief plus a one-line summary you can paste into any image generator. Splitting the idea into labelled fields forces every detail to be stated, so the model has less room to guess.

Worked in our testIt followed the format: valid JSON, the four required keys plus fixed_details, and a single-sentence "Prompt:" line. The era, mood and light are applied consistently, and the light direction is coherent. Some field values are long rather than short. Avoiding "extra text" sits awkwardly beside the handwritten chalkboard lettering, which has no specified wording. The "Closed" sign was invented and then fixed as exact text.

The prompt

You are an art director who writes precise briefs for AI image generators.

Task: Turn my rough idea into a structured JSON image brief that an image model can follow without guessing.

Rough idea: {rough_idea}
Era and mood: {era_and_mood}
Camera and lighting: {camera_and_lighting}

Steps:
1. Split the idea into four top-level keys: subject, photography, background and avoid.
2. Inside each key, add short nested fields such as clothing, lens, props or light source. Fill each with one concrete visual detail instead of a vague adjective.
3. Apply the era, mood, camera and lighting details consistently across every key.
4. Add a fixed_details list naming anything that must stay exactly as described, such as colours, shapes or printed text.
5. Check the brief for contradictions and fix them before you answer.

Output: Valid JSON only, with no commentary. After the JSON, add one final line beginning "Prompt:" that sums up the brief in a single sentence I can paste into an image tool.

Change the parts in curly braces. Use it when your image results keep drifting from what you pictured, or when you need many images in the same consistent style.

What happened when we ran it

We sent

You are an art director who writes precise briefs for AI image generators.

Task: Turn my rough idea into a structured JSON image brief that an image model can follow without guessing.

Rough idea: A small bakery owner stands behind her counter holding a tray of cinnamon buns. The shop is cosy and a little cluttered, with handwritten price cards and a chalkboard menu. I want it to look like an authentic photo for our website homepage, not a stock image.
Era and mood: Present day, warm, friendly and unfussy, early morning before opening time
Camera and lighting: 35mm lens, shallow depth of field, soft window light from the left, slight film grain, shot at eye level

Steps:
1. Split the idea into four top-level keys: subject, photography, background and avoid.
2. Inside each key, add short nested fields such as clothing, lens, props or light source. Fill each with one concrete visual detail instead of a vague adjective.
3. Apply the era, mood, camera and lighting details consistently across every key.
4. Add a fixed_details list naming anything that must stay exactly as described, such as colours, shapes or printed text.
5. Check the brief for contradictions and fix them before you answer.

Output: Valid JSON only, with no commentary. After the JSON, add one final line beginning "Prompt:" that sums up the brief in a single sentence I can paste into an image tool.

The model replied

{ "subject": { "person": "woman in her late 40s, the bakery owner, relaxed candid expression with a small genuine smile, looking slightly off-camera toward the left window", "pose": "standing behind the counter, both hands holding the tray at waist height, shoulders relaxed", "clothing": "natural linen apron in oatmeal beige over a rolled-sleeve navy shirt, light dusting of flour on the apron", "hair": "dark brown hair tied back loosely with a few loose strands", "held_item": "rectangular metal baking tray of 8 freshly baked cinnamon buns with white icing drizzle, slightly uneven and golden-brown", "skin_and_texture": "natural skin texture with no retouching, faint flour smudge on one forearm" }, "photography": { "style": "candid documentary-style photo, authentic small-business homepage image", "lens": "35mm", "depth_of_field": "shallow, with focus on the owner's face and the buns, and the background softly blurred", "camera_height": "eye level with the subject", "framing": "medium shot from mid-thigh up, subject slightly right of centre, tray in the lower third", "light_source": "soft natural window light coming from the left of frame", "light_quality": "gentle early-morning daylight with soft shadows falling to the right and a warm tone", "texture": "slight film grain, natural colour, no heavy filters", "time_and_mood": "present day, early morning before opening, warm, friendly, unfussy" }, "background": { "setting": "small independent bakery interior, cosy and a little cluttered", "counter": "worn light-wood counter with a glass cake stand, a paper bag stack and a small brass till", "chalkboard": "black chalkboard menu on the back wall with white handwritten chalk lettering, softly out of focus", "price_cards": "small handwritten cream cards in black marker propped on the counter edge, one reading 'Cinnamon Buns'", "clutter": "stacked flour sacks, a wooden rack of bread loaves and a small potted herb on the shelf", "window": "large shop window on the left with a flipped door sign reading 'Closed' just visible, daylight spilling in", "colour_palette": "warm creams, honey wood, navy and chalk-black with golden-brown pastry highlights" }, "avoid": { "look": ["stock-photo posing", "overly perfect toothy smile", "symmetrical staged composition", "glossy over-retouched skin", "HDR or oversaturated colour"], "lighting": ["harsh flash", "cold blue light", "studio softbox look", "light coming from the right"], "content": ["customers in shop", "other staff", "brand logos", "modern digital screens", "empty or burnt trays"], "technical": ["garbled or extra text", "distorted hands or extra fingers", "deep focus with sharp background", "wide-angle distortion", "heavy vignette"] }, "fixed_details": [ "Cinnamon buns have white icing drizzle and a golden-brown colour", "Tray is rectangular and metal", "Apron is oatmeal beige over a navy shirt", "Chalkboard is black with white handwritten chalk lettering", "Price card reads exactly 'Cinnamon Buns' in black marker on cream card", "Door sign reads exactly 'Closed'", "Window light comes from the left only", "35mm lens at eye level with shallow depth of field and slight film grain", "Setting is early morning before opening, present day" ] } Prompt: A candid, authentic 35mm eye-level photo with shallow depth of field and slight film grain, showing a woman bakery owner in an oatmeal linen apron and navy shirt holding a metal tray of iced, golden cinnamon buns behind a worn wooden counter in a cosy, slightly cluttered shop with handwritten price cards (one reading "Cinnamon Buns"), a blurred chalkboard menu and a "Closed" sign, lit by soft warm early-morning window light from the left, with no stock-photo posing, no logos and no garbled text.

Run on Claude Sonnet 5.5, 29 Oct 2026. Automatic check: 4 out of 5.

Mentioned for

Gemini, Nano Banana Pro, ChatGPT, Claude

Where it came from

Found on GitHub, shared by ZeroLu: ZeroLu/awesome-nanobanana-pro.

This is our own version of the idea, written and tested by us. The original may differ.

More images and video prompts

Images and video Worked in our test
Break Image Prompts Into Labelled Parts

You get a rough image prompt sorted into labelled parts, with clashes flagged and a tidy rebuilt version. Seeing each phrase's job makes it easier to spot padding and contradictions, such as asking for both watercolour and photorealistic.

Found on GitHub, shared by Moonvy
Images and video Not tested yet
Four-Part Photo Edit Brief

You get an edited version of your own photo in a new setting, with your face and the number of people left alone. Splitting the brief into subject, setting, lighting and a clear 'keep the same' rule gives the tool less room to drift.

Found on Web, shared by Atlas Cloud
Images and video Partly worked
Rewrite Video Prompts Into Six Parts

You get a clean six-part breakdown of your old video prompt, plus a single polished version ready to paste into your video tool. It works because each element has its own clause, so nothing is blurred together and gaps are filled openly rather than guessed silently.

Found on Web, shared by Nadia Dubois (Tech Insider)