AI, Photo & Developer Tools

Repairing and enlarging photographs: the order matters

Three apps here work on old or low-resolution images, and using them in the wrong order produces the plastic, over-smoothed look people associate with AI photos.

Old Photo Restorer goes first. It heals scratches, tears and stains, and runs a dedicated model over each face - up to 16 in one photograph - recovering the features that are there rather than inventing new ones. Only then does SharpUpscale enlarge the result up to 4x. Upscaling first simply magnifies the damage: the model has no way to distinguish a scratch from a feature, so it sharpens both. That failure mode is the subject of why AI-upscaled photos look fake, and the practical walkthrough is upscaling old family photos.

If the destination is paper rather than a screen, the requirements change - see preparing photos for print and JPEG vs PNG for upscaling, since starting from a re-compressed JPEG limits what any model can recover.

Changing a photo on purpose, and hiding what should not be in it

Photo Cartoonizer is the deliberate-transformation one - turning a photograph into illustration or painterly styles. What it is doing under the hood is explained in what neural style transfer is, and the two common jobs are an anime profile picture and making a photo look like a painting.

Face Blur AI is the opposite: it finds faces, licence plates and other identifying detail and blurs them before a photo or video is published. That matters most for other people's children and for photos taken in public - blurring faces before posting and sharing photos of children safely. It is also listed under Business & Documents, where the same job is a publication step rather than a personal one.

HEIC to JPG Batch Converter solves the format problem iPhones created: a photo library that other software cannot open. See iPhone photo formats explained, HEIC vs JPEG, and converting without losing quality. AI Face Score is the entertainment one - what it does and does not measure is set out in what face analysis actually measures.

Text and pattern tools

Two developer-adjacent tools sit in this category because they solve the same shape of problem: text arrives in the wrong form and has to be reshaped before it is useful.

Regex Tool is a pattern editor with live highlighting, capture groups and reusable snippets, so a pattern is built by watching it match rather than by guessing. Start from the regex cheat sheet, avoid the common regex mistakes, and see real-world tips or the worked example of extracting e-mail addresses from text.

TextLab covers the jobs that do not deserve a script: cleaning up messy text, rearranging lines, blocks and sentences, formatting JSON, and the encoding problems everyone eventually hits - with Base64 and delimiters explained separately.