Link copied!

AI Open Eyes Fix

Upload a photo where someone blinked, and AI reconstructs naturally open eyes that match the person's iris color and eyelid shape, so the shot is no longer lost.

Drop your photo with closed eyes here or browse

JPG, PNG, WebP (max 20MB)

Opening eyes...

AI is naturally opening the closed eyes in your photo

Result
Download

How It Works

  1. Upload a photo where someone has their eyes closed
  2. AI automatically detects and opens closed eyes
  3. Download your fixed photo with natural-looking open eyes

Tips

  • Works best on photos where eyes are only slightly closed or blinking
  • Clear, well-lit face photos produce the most natural results
  • Front-facing photos with visible eyelids work best
  • The AI matches the person's natural eye color and shape

Also Try

Why Use AI Open Eyes Fix

Eye-Specific Reconstruction

The model targets only the eye region, reading iris color, eyelid shape, and lash angle from the rest of the face before generating open eyes that look like they belong there.

No Manual Masking

You do not draw a selection or trace eyelids. Upload the photo, and the AI locates the closed eyes, inpaints them, and outputs the result without any extra steps from you.

Rescue the Blink Shot

When one person blinked in an otherwise perfect group photo, you no longer need a retake or hours in Photoshop transplanting eyes from another frame.

Rest of Photo Unchanged

Skin tone, hair, clothing, and background are left exactly as they were. Only the eye area changes, so the fixed photo does not look touched up anywhere else.

Perfect For

Group Photos Selfie Fixes Event Photography Wedding Photos Family Portraits ID Photos

Powered by AI Eye Reconstruction

FLUX.1 Eye Reconstruction Engine

Open Eyes Fix runs on FLUX.1 dev in image-to-image mode. Rather than pasting a generic eye template, the model reads the full face context, including skin undertone, brow arch, and available eyelid detail, and generates open eyes that are consistent with that specific face.

The inpainting mask covers only the eye region. Everything outside that area, including skin, hair, glasses frames, and background, is passed through untouched. Inside the mask, the model reconstructs eyelid position, lash direction, iris color, and pupil size to produce a result that looks like a photo taken a fraction of a second later rather than an edited image.

Frequently Asked Questions

The model reads the whole face before touching the eye area. Brow shape, visible eyelid edges, skin tone, and the iris color visible through a partial blink all give it enough information to reconstruct open eyes that fit that particular face. It is not inserting a generic eye; it is generating one that belongs there.

Yes. Fully closed eyes produce good results in most cases, though a mid-blink photo where part of the iris is still showing gives the model more to work with and tends to look slightly more natural. Either way is worth trying.

One face at a time gives the best results. If several people blinked, crop each face individually, run the fix, and then reassemble. Processing a dense group at full frame sometimes works but the per-face reconstruction quality is lower.

No. The inpainting mask is limited to the eye region. Skin, hair, clothing, background, and all other facial features are passed through from the original without any modification.

Thin wire frames and clear lenses work reasonably well. Thick plastic frames or strong tints block the iris information the model depends on, so results are less reliable in those cases.

Most photos complete in 15 to 30 seconds. Larger files take a moment longer.

It can, but sleeping faces have relaxed muscles and softer skin around the eyes, so reconstructed open eyes may look slightly incongruous with the rest of the expression. The tool was built for blink-moment photos where the face is otherwise alert.

JPG, PNG, and WebP, up to 20MB. The corrected photo is delivered as a PNG.

No. This tool opens closed eyes; it does not correct red-eye from flash. Because the eyes are fully reconstructed rather than recolored, the result will not have red-eye artifacts, but red-eye correction in otherwise open-eyed photos is a separate problem.

The model infers iris color from whatever part of the eye is visible in the original, whether a sliver of iris showing through a blink or the skin tone context around the eye. For most photos the match is close. If a mid-blink shot shows no iris at all, the model makes its best inference from surrounding facial features.

AI Open Eyes Fix vs Other Methods

Feature Luxoret AI Manual / Traditional Other Tools
Cost per Use $0.09 $50-$200+ per project $0.15-$0.30 per image
Speed Results in seconds Hours with Photoshop Minutes per image
Skill Required None — fully automated Professional editing skills Some learning curve
Quality AI-powered, photorealistic Depends on editor skill Template-based, limited
Accessibility Browser-based, any device Desktop software required App or desktop required
Revisions Unlimited instant retries Time-consuming redo Limited by credits