- Home
- Photo tools
- AI Relight: Explore New Lighting in Your Photographs
Give the existing scene a different light
AI relight
Use AI relight to explore how a photograph might look with softer window light, a warmer afternoon tone, or a more directional studio treatment. Describe where the light comes from and what should remain recognizable. The result is a generated lighting interpretation, so inspect the subject as carefully as the new highlights.
Your original stays separate · Review the price before generating
Basic edits are free. AI images start at 10 credits with these settings. Review the price on the canvas before generating.
Before you start
AI relight uses a written lighting brief to generate a new version of your photograph. Specify direction, softness, and warmth, then compare shadows, reflections, and subject details with the original. It is not a dedicated physical lighting simulator and cannot reliably recover detail lost in clipped highlights.
Choose the lighting problem first
Start an AI relight request with the issue you can see. A face may feel flat, a room may appear cold, or a product may blend into a dark setting. These problems need different lighting descriptions, even when all three could be described loosely as making the image brighter.
Decide whether you want a gentle treatment or a deliberate mood change. Softening the apparent window light is a smaller request than turning a midday room into a neon scene. A stronger reinterpretation gives you more details to inspect because it can affect materials, color, and background appearance together.
Write an AI relight brief you can inspect
An AI relight brief should identify where the light originates relative to the image. Left, right, above, or behind the subject is easier to evaluate than cinematic lighting. Add whether shadows should be soft or defined, then describe the color temperature in everyday terms such as neutral or warm.
Mention important surfaces that should remain readable. A dark jacket may need visible folds, while a glass bottle needs a plausible highlight without losing its outline. These instructions guide a general image model. They do not expose a virtual light rig or promise a measured, physically exact result.
| Desired effect | Illustrative lighting brief | Main review point |
|---|---|---|
| Gentle portrait | Soft neutral window light from the left | Face shape and eye reflections |
| Warm interior | Late-afternoon warmth with readable shadows | Wall color and existing room layout |
| Product separation | Subtle light along the rear edge | Accurate product silhouette |
| Dramatic concept | Defined side light with restrained fill | Consistent shadows and retained detail |
Use a source with recoverable visual information
A good AI relight starting image has enough visible detail in both light and dark areas to compare afterward. Completely white highlights do not contain reliable texture to preserve. Deep black regions may leave the model to invent what was there.
Choose the original file when possible instead of a heavily filtered export. Supported inputs are JPG, PNG, WebP, and HEIC up to 20 MB. If a reflection or shadow is central to your composition, include it in the frame and mention it in the brief. Cropping it away before generation removes useful context for judging consistency.
Make an AI relight comparison
Save a short lighting description before starting. It should be specific enough that another person can tell whether the image meets the request.
Identify the existing light
Look at the nose shadow, window highlights, or the shadow under an object. These clues tell you which parts need to change together if the light direction changes.
Request one treatment
Describe light direction, softness, and warmth. Preserve the subject, materials, and arrangement explicitly, and avoid adding unrelated style or object changes.
Inspect the brightest surfaces
Compare skin, polished surfaces, and white fabric. Watch for newly invented texture or clipped patches that erase information you wanted to keep.
Follow the shadows
Check cast shadows and contact shadows across the frame. A pleasing highlight is insufficient if the ground shadow still points toward an incompatible light source.
Reflections reveal an inconsistent AI relight
AI relight can change the appearance of a light source without making every reflective surface agree. Check glasses, windows, water, and polished products. A soft highlight on a face may conflict with a sharp reflection in the glasses if the rest of the scene has not followed the request.
Also inspect contact shadows, such as the dark area directly under a cup or shoe. These help objects feel attached to their surfaces. If a lighting change removes them entirely, the subject can appear to float. Review the image at ordinary size after the close inspection, because a globally inconsistent mood is easier to see when you step back.
Keep lighting separate from identity and product changes
A useful AI relight concept still needs a recognizable subject. On a portrait, compare facial proportions, expression, and skin texture. On a product, check labels, edges, and material appearance. A warmer image can make a surface look like a different finish, which may be unsuitable for a listing.
When appearance accuracy matters, use the original as your reference throughout the review. Do not assume that a light-only request preserves exact geometry. If you only need a new background or a different sky, use that focused workflow and keep the intended lighting changes explicit.
Revise the light in small steps
If the first AI relight version is too orange, ask for less warmth while retaining the same direction and softness. If AI relight makes the shadows too deep, ask for more visible detail there. One adjustment at a time lets you identify what improved.
Return to the original when the subject has drifted. Repeatedly editing a changed version can carry altered features into later results. There is no dedicated batch lighting workflow here, so evaluate each photograph separately if you are building a series. Matching the prompt does not establish identical lighting across different source scenes.
For a tabletop scene, compare the underside of the object as well as its bright face. AI relight should not make a heavy object appear suspended above its surface. If the contact shadow weakens, revise that specific relationship before increasing contrast elsewhere. A small grounded shadow can matter more to the result than a dramatic highlight across the top.
Choose a useful output size
AI relight runs through general image generation. A 1K result costs 10 credits, 2K costs 20, and 4K costs 30. Every new lighting variation is paid, while internal retries for the same job do not add another charge. Select the output size before submitting.
Download PNG, JPG, or WebP once the highlights and shadows pass your review. Keep both the original and the edited file. Free crop and resize tools can prepare the chosen version for its destination without another generation. A larger image may reveal more detail for inspection, but it does not guarantee physically accurate lighting.
Example request
Relight this image with soft warm light from the left. Preserve the subject, facial features, materials and room layout. Keep shadow detail visible and reflections consistent.
Adapt the subject and details to your image. This is a starting instruction, not a sample result.
Questions about this edit
Is AI relight a brightness slider?
No. AI relight asks the image model to interpret a lighting description and generate a new version. That can affect more than overall brightness, including shadows, material appearance, and details you asked to preserve.
Can AI relight recover blown highlights?
It may invent plausible appearance in a bright region, but it cannot reliably recover the original information if it is absent from the source. Treat newly generated texture as an interpretation, especially when accuracy matters.
Can I place individual studio lights?
This page offers prompt-based lighting exploration rather than a dedicated light placement interface. Describe the desired direction and softness in words, then inspect the result. There is no promise of measured light intensity or physical simulation.
Why do some reflections still look wrong?
The model may not change every related surface coherently. Check glasses, windows, and polished objects against the intended light direction. A narrower AI relight request or a clearer source may give a more useful comparison, but consistency still needs review.
Can AI relight keep a product's exact color?
Color appearance can change with the requested lighting, and generation can introduce additional drift. Compare the output with accurate product references. Use actual product photography where a customer needs dependable evidence of color or finish.
Explore a lighting treatment
Bring your image and a clear request. Choose the output settings and review the task price before generating.