10 ChatGPT Photo Editing Prompts You Can Copy and Paste

/identitylock

Edit the uploaded photo to improve its exposure, colour balance, contrast and overall photographic quality while prioritising the identity of every person pictured.

Preserve facial proportions, eye shape and colour, nose, lips, jawline, hairline, hairstyle, age, skin tone, expression and body shape. Keep natural skin texture and individual characteristics. Do not apply beauty filters, reshape features, slim bodies or make anyone look younger.

Keep the clothing, pose, background, camera angle and framing unchanged. Make restrained photographic improvements rather than redesigning the image. If an enhancement would require inventing facial detail, leave that area closer to the original.

Return one edited image with the original aspect ratio and no added text.

/studio-quality

Edit the uploaded photo to look professionally lit while keeping the original person, setting and composition.

Create the appearance of a large soft key light positioned slightly above and to one side of the camera, gentle fill light to soften deep facial shadows, and subtle separation between the subject and background. Correct poor exposure and unwanted colour casts while keeping skin tones believable.

Preserve the person's identity, expression, pose, clothing and body proportions. Retain pores, hair detail and fabric texture. Keep highlights controlled and shadows dimensional. Avoid plastic skin, excessive sharpening, artificial HDR or an obvious beauty filter.

Keep the existing background rather than replacing it with a studio backdrop. Return one edited image in the original aspect ratio without text.

/upscale

Enhance the apparent clarity of the uploaded photo while staying as close as possible to its original content.

Reduce visible compression artefacts, distracting noise and mild softness. Improve edges and textures conservatively, especially around eyes, hair, clothing and important objects. Preserve the original faces, expressions, colours, lighting, framing and background.

Do not invent text, facial features, patterns or objects that cannot be confidently inferred. Keep ambiguous details subtle. Avoid sharpening halos, exaggerated pores, waxy skin and oversaturated colours.

Return the enhanced image at the highest available output resolution, keeping the original aspect ratio. Do not add captions or comparison panels.

/clean-up

Clean up distracting clutter in the uploaded photo while preserving its main subject and original setting.

Remove obvious temporary distractions such as scattered rubbish, loose cables, discarded packaging and untidy piles in the background. Preserve furniture, architecture, intentional decor and anything the subject is wearing, holding or using. If an object's importance is ambiguous, keep it.

Reconstruct the exposed surfaces with consistent texture, perspective, lighting and shadows. Preserve all people, their faces, poses and clothing. Keep the camera angle, crop and depth of field unchanged.

Make the scene look naturally tidy, not empty or redesigned. Return one edited image with no text.

/pose-only

Change only the person's pose in the uploaded photo.

New pose: [describe the pose—for example, turn the torso slightly sideways while keeping the face looking towards the camera, with relaxed shoulders and arms].

Preserve the person's identity, facial proportions, age, skin tone, hairstyle, expression, body shape and clothing. Keep the same background, camera position, lighting, crop and overall photographic style.

Adjust clothing folds and shadows only where the new pose requires it. Keep shoulders, elbows, wrists, hands and legs anatomically plausible. Do not add props, accessories or extra limbs. Avoid extreme twisting or exaggerated posture.

Return one edited image in the original aspect ratio with no added text.

/remove-object

Remove only this element from the uploaded photo: [describe the object or background person and its location].

Keep everything else unchanged, including the main subject's face, body, clothing, pose and position. Do not crop the image to hide the unwanted element.

Reconstruct whatever should appear behind the removed element using the surrounding scene as reference. Match perspective, texture, lighting, depth of field and shadows. Remove any reflection or shadow belonging solely to the removed element, while preserving those belonging to other subjects.

Avoid repeated textures, smudged patches and visible editing seams. Return one edited image with the original dimensions or aspect ratio and no text.

/background-swap

Replace the background of the uploaded photo with: [describe the new setting].

Preserve the foreground subject's identity, facial features, expression, hairstyle, clothing, body proportions, pose and scale. Keep fine edges such as hair, glasses and fabric clean and natural.

Make the new setting match the original camera height, perspective and depth of field. Match its lighting direction and colour temperature to the subject wherever possible. Add realistic contact shadows and reflections where needed so the subject feels physically present.

Do not redesign or beautify the subject. Avoid artificial cutout edges, halos and inconsistent scale. Return one cohesive photorealistic image in the original aspect ratio without text.

/golden-hour

Relight the uploaded photo as if it were taken during natural golden hour.

Add warm, low-angle sunlight from one consistent direction, gentle highlights, soft ambient fill and believable shadows. Keep skin tones natural, preserve detail in bright areas and avoid covering the image with a uniform orange filter.

Preserve the identity, expression, pose, clothing and body proportions of every person. Keep the original setting, objects, camera angle and framing. Make only the lighting and colour changes needed for the effect.

If the sky is visible, warm it subtly without adding a dramatic replacement sunset. Avoid fake lens flare, glowing outlines and excessive saturation. Return one edited image without text.

/outfit-swap

Replace the person's outfit in the uploaded photo with: [describe the garments, colours, materials and fit].

Preserve their face, hairstyle, skin tone, expression, body shape, pose and position. Keep the background, camera angle and lighting unchanged. Do not alter their physique to fit the outfit.

Make the new clothing follow the existing body and pose naturally, with believable seams, folds, fabric weight, shadows and layering. Preserve hands and any objects being held. Keep existing accessories unless they directly conflict with the requested clothing.

Do not invent brand logos or change unrelated parts of the image. Return one realistic edited photo in the original aspect ratio with no text.

/expand-frame

Expand the canvas of the uploaded photo to [target aspect ratio, such as 9:16 for Stories or 16:9 for a website banner].

Preserve the original image area and keep the subject's face, body, pose, clothing and proportions unchanged. Add space primarily around the subject rather than enlarging or stretching them.

Extend the existing environment naturally, matching its perspective, architecture, textures, lighting, shadows and depth of field. Do not duplicate people, objects or repeating background features.

Place the subject [centrally / on the left / on the right], leaving useful empty space in [location]. Where the original photo cuts off body parts, prefer extending the surrounding background rather than inventing anatomy.

Return one expanded image without text, borders or watermarks.