AI Photo Fixer

Common scenarios

• A birthday candid where the subject moved just before the shutter — faces are smeared with motion blur
• A phone shot in low light where camera shake made the entire frame soft and indistinct
• A portrait where autofocus locked onto the background instead of the subject's face

Best practices

• Describe the blur type for better results: 'motion blur' (subject moved), 'camera shake' (whole image blurs), or 'out of focus' (soft but no directional smear) each respond to slightly different sharpening approaches
• For portraits, specify which element to prioritize: 'sharpen the faces and eyes while keeping the soft background' produces cleaner results than a generic sharpen
• On phone photos with heavy noise added to a blurry image, ask to 'sharpen and remove grain' in one prompt — fixing blur without addressing noise often looks over-processed
• If the first pass over-sharpens and creates halos, follow up with 'reduce the halo artifacts around the edges while keeping the sharpness improvement'

Sample prompts

Common scenarios

• A restaurant or indoor event shot taken without flash — the camera boosted ISO and the image looks speckled and muddy
• A sports action photo shot at high shutter speed in dim indoor lighting — every surface is covered in color noise
• A scan of an old film photo where the grain from the original film is now amplified by the scanning process

Best practices

• Distinguish between film grain and digital noise in your prompt: 'reduce the digital noise' for phone/camera shots, 'soften the film grain while keeping the photo's vintage character' for scanned old prints
• Specify if you want to preserve texture: 'remove the noise while keeping the skin texture natural and the fabric detail intact' prevents the AI from smoothing everything into plastic-looking skin
• For color noise (where individual pixels show random red, green, blue specks), add 'fix the color noise and restore natural colors' — color noise needs different treatment than luminance grain
• After denoising, follow up with a light sharpening pass if details look too soft: 'sharpen the edges and fine details without bringing back the grain'

Sample prompts

Common scenarios

• A grandparent's wedding portrait from the 1960s with a diagonal scratch across the faces, heavy yellowing, and areas of fading around the edges
• A childhood photo that got wet — the image has tide marks, color bleeding, and sections where the emulsion has lifted
• A photo stored face-down in an album for decades — the backing has stuck to the surface and pulled away chunks of the image

Best practices

• List every damage type you can see in one prompt: 'fix the scratch across the center, repair the yellowing, restore the faded right edge, and remove the water stain in the lower left' — the AI addresses each issue in a single pass
• For photos with faces, always include 'preserve the original faces exactly — repair damage without altering anyone's appearance' to prevent the AI from changing who someone looks like while reconstructing damaged areas
• Work in stages for severely damaged photos: fix structural damage (tears, missing sections) in the first pass, then address color issues (yellowing, fading) in a follow-up pass on the repaired result
• Scan at 300–600 DPI before uploading for best results — a phone photo of a print works but a proper scan gives the AI much more detail to work with

Sample prompts

Common scenarios

• An outdoor photo on a bright day where the background is properly exposed but the subject's face is a dark silhouette against the bright sky
• A flash photo at a party where every face looks flat, shiny, and washed out — all depth and natural skin texture gone
• A night shot where the camera exposed for the bright lights, leaving the rest of the scene nearly black

Best practices

• Name the specific problem: 'fix the backlit silhouette and recover the face detail', 'reduce the harsh flash and restore natural skin texture', or 'brighten the underexposed shadows while keeping the highlights' each get more targeted results than 'fix the lighting'
• For backlit subjects, tell the AI what to prioritize: 'bring out the face detail while keeping the sky natural' prevents it from blowing out the background in the process of brightening the subject
• For flash photos, 'soften the harsh shadows under the nose and chin, reduce the shine on skin, and restore depth' is a complete instruction that addresses the three main flash problems at once
• After fixing exposure, check if colors need a separate pass: exposure correction sometimes shifts color balance, and a follow-up 'correct the color cast and restore natural skin tones' keeps results accurate

Sample prompts