FAQ
Why do faces look stretched near the edges of a photo?
A wide-angle camera fits a large scene into a flat image. As the viewing angle grows, objects near the edges are spread across more pixels, so a face there can look wider than the same face near the centre. This is a normal result of perspective projection, not a flaw in the person.
How does the correction work?
It is geometry, not generative editing. Facelign finds faces, reads the camera’s field of view when available, and builds a smooth local perspective correction around faces that need it. The photo is warped once, without generating or replacing facial details.
Is Facelign really free?
Yes. It is free to use and needs no account. File-size, image-size, daily-capacity, and rate limits keep the public service affordable, so it may occasionally ask you to try again later.
How do you use my photos, and how long do you keep them?
Your photo is used only to perform the correction. The original is deleted after processing finishes; generated results are available through short-lived private links and have a one-day automatic deletion rule as a safety net. Photos are not used for training, advertising, or identification. See the Privacy page for the full explanation.
How do I keep the original filename and photo tags on Android?
Install Facelign and send photos to it from Android’s Share menu. This is the supported Android workflow for retaining the source filename and the metadata Facelign can preserve.
- Open facelign.com in Chrome.
- Open the Chrome menu and choose Install app or Add to Home screen.
- Open the original photo in Photos, Gallery, or Files.
- Tap Share, then choose Facelign. Correction begins automatically.
If you choose a photo from facelign.com in a regular browser instead, Android may provide a renamed, metadata-reduced copy.
Which photo formats are supported?
Facelign accepts JPEG, PNG, WebP, HEIC, and HEIF images. Corrected JPEG input remains JPEG; HEIC, HEIF, PNG, and WebP corrections are delivered as JPEG.
Is the original image quality preserved?
The full pixel dimensions are preserved. Corrected images are encoded as high-quality JPEGs, so they are not bit-for-bit or lossless copies. HEIC, HEIF, PNG, and WebP inputs are converted to JPEG, and special HDR or motion data is not retained.
Are photo tags preserved?
Facelign preserves EXIF information, including GPS when present, and the colour profile in corrected JPEGs when the source format provides them. Photos are saved the right way up, even when the original used a rotation tag. XMP and format-specific auxiliary metadata are not guaranteed.
Why are some tags missing after correction?
Sometimes the browser, photo provider, cloud gallery, or another app supplies a privacy-filtered or exported copy instead of the original file. Facelign cannot restore metadata it never receives. On Android, install Facelign and share the original from Photos, Gallery, or Files. On other platforms, select the original file rather than a copy previously exported through social media.
Why does a mobile upload sometimes have a strange filename?
The browser or photo provider chooses the filename it gives to a website and may substitute a generated name for privacy or compatibility reasons. Facelign preserves the name it receives as the basis for the corrected filename, but cannot recover a name that was not provided. On Android, the installed Facelign share target is the supported way to retain the source filename.
Why does the service sometimes say it is busy?
Facelign uses a deliberately limited amount of computing capacity so the service can remain free and resist automated abuse. Your browser retries short bursts automatically. If demand stays high or the daily allowance has been reached, wait a little and try again later.
Why do I see a message about missing field-of-view information?
The correction normally uses the 35 mm-equivalent focal length stored in the photo’s EXIF data. Older photos, screenshots, edited exports, and images downloaded from social media may not contain it. Facelign then uses a conservative default field of view of 80° and tells you that it has done so.
Why can results be imperfect when a face touches the edge?
Correction needs source pixels on both sides of a face so it can redistribute them naturally. At the edge of the frame, some of those pixels were never captured. Facelign does not invent missing facial or background content, so an extreme edge face can remain imperfect or show stretching near the boundary.
What are the known limitations?
Facelign corrects the primary still image only. Live Photos, Android Motion Photos, HEIC sequences, depth maps, gain maps, and Ultra HDR effects are not preserved. HEIC output is not currently supported. Images are limited to 25 MB and 32 megapixels, and automatic detection may miss small, obscured, or unusual faces.
I have an idea or question. How can I get in touch?
Contact details are on the About page.