Make Noisy, Blurry Photos Usable Again With a Local AI Image Enhancer

8 hours ago 7

Every photographer has a folder of images they love but never use, because the file is too noisy, a little soft, too small to print, or decades old. Aiarty Image Enhancer is built to rescue those exact shots, and it does all of the work on your own computer.

Most of the AI software we cover on Fstoppers is about speed across an entire shoot. Tools like Imagen and Aftershoot apply your editing look to thousands of wedding or event photos at once. Aiarty Image Enhancer works more like a repair kit for individual frames. It removes noise, sharpens soft shots, enlarges small files, erases distractions, and brings faded family snapshots back to life.

Aiarty is also running a Black Friday deal for our readers that makes it a pretty easy add to your toolkit. With the code FSTOPPERS, a lifetime license drops to $79 from $109, and one license covers up to three computers. You can grab the Black Friday offer now as it's only valid through December 4th, and I'll cover the full details at the end of this article.

How Aiarty Runs on Your Computer

A lot of newer AI apps upload your photos to a server and charge tokens or credits for every edit, so that was the first thing I checked with Aiarty. The software runs locally on your Windows PC or Mac, and your images never leave your hard drive. The only time it needs the internet is when you activate your license and download each AI model the first time you use it. After that, you can unplug from the web and edit anywhere, which is great if you work on a laptop while traveling.

Aiarty shows the original and the enhanced file side by side. This is the snowy village example from later in this article, zoomed to 354% and run through More-Detail GAN v3 at x2.

There are no tokens, credits, or monthly limits to keep track of. Aiarty says the lifetime license includes unlimited local processing, so a batch of 500 photos costs the same as a single image. The tradeoff is that your computer does the heavy lifting. Aiarty needs 64-bit Windows 10 (version 1809 or later) or macOS. It uses your graphics card to speed things up, although it can also run on the CPU alone if you're patient. My editing PC still runs an older NVIDIA GeForce GTX 1070, and it handled every example in this article. Keep in mind that bigger upscales take longer on older hardware.

The Models and Settings

Aiarty gives you a handful of AI models to choose from, and each one has its own personality. Real-Photo v3 is the conservative option, designed to stay faithful to the original photo, which makes it a good choice for portraits and wildlife. More-Detail GAN v3 is more aggressive and builds out stronger texture and sharpness. It's the model Aiarty suggests for landscapes, cityscapes, blurry images, and old photos. AIGCsmooth v3 is tuned for AI-generated art and images with smooth surfaces. The menu also includes the older More-Detail GAN v2 and a Smooth Diff v2 model for illustrations.

The AI Model menu in Aiarty Image Enhancer. The 2-Pass option underneath gets its own model setting, which I left on More-Detail GAN v3 for every example.

After choosing a model, you pick how much to enlarge the file. The x1 setting cleans up the image without changing its pixel dimensions, and you can also go to x2, x4, x8, or a specific target like 4K. A Strength slider controls how hard the model works, and an optional 2-Pass setting runs a second cleanup on very grainy files. Face Restoration and the AI Eraser can be switched on as extras. If you want a starting point for each type of photo, Aiarty has a useful AI model and settings guide. The company also publishes a set of free test pictures with suggested settings for each one. I used that collection for the examples below, so you can download the same files and compare your results with mine.

Noise Reduction

Noise is the most common reason a good photo stays in the reject pile. Blue hour landscapes, dark receptions, and indoor sports all push your ISO into territory where the shadows turn to mush. Raw files give you plenty of options in your main editor, while a JPEG you exported years ago or an image someone texted you is much harder to clean up.

The full original frame. The slider below zooms into the cabin.

Wildlife is where noise reduction earns its keep. Birds and animals are usually shot at fast shutter speeds, long focal lengths, and high ISOs, and then cropped heavily. For this kind of image, I used Real-Photo v3, since it's the model designed to stay closest to the original feathers and fur. Aiarty has a whole page on how to enhance wildlife photos if that's the kind of shooting you do.

Soft and Blurry Photos

Missed focus is the other heartbreaker. If you shoot kids, pets, or events, you've definitely had the perfect expression land on a frame where the focus fell on an ear or the background. Traditional sharpening only adds contrast along edges that are already there, which is pretty limiting when the whole face is soft. More-Detail GAN v3 takes a different approach, since it reconstructs texture in the eyes, hair, and fabric.

The full original frame. The slider below zooms into his face.

Bigger Files for Print

Maybe you cropped deep into a frame or shot it on an older 12-megapixel camera. Maybe a client wants a wall print from a file that was only meant for Instagram. Either way, you need more pixels than you have. Aiarty can enlarge an image up to 32K on Windows and 16K on a Mac. It saves TIFF, PNG, JPG, and DNG files, so the bigger version can go straight back into your regular editor.

The full original frame. The slider below zooms into her face.

For the black-and-white aerial of a city below, I ran the 2,100-pixel original at x2, which doubled it to 4,200 pixels wide. If you need more for a big wall print, the same model goes up to x8, which would take this file past 16,000 pixels wide.

The full original frame. The slider below zooms into the center buildings.

The AI Eraser

With the AI Eraser, you paint over a stray person, an object, some text, or a scratch, and Aiarty fills in the area behind it. In the example below, I used it on the people walking across the scene, along with their long shadows, and upscaled the tiny original in the same pass. Doing this in Photoshop, of course, means working back and forth between Content-Aware Fill and the Clone Stamp. Aiarty handles the cleanup and the enlargement together.

Each blue stroke marks a person and their shadow for the AI Eraser, which the app still labels as experimental.

Old Family Photos

Old photos are a big reason to own a tool like this, especially around the holidays. Most of us have a shoebox of faded prints or a hard drive of low-resolution scans from our parents and grandparents. Restoring those by hand in Photoshop is slow, skilled work, and it's often the kind of job people pay a retoucher to do. Aiarty combines enlargement, noise cleanup, Face Restoration, and color correction in one pass. That makes it super approachable for someone who just wants to print a family photo for a relative.

Face Restoration has two modes. Fidelity stays close to the original face, while Rebuild reconstructs more of it, and you can set its strength separately from the rest of the image. For family photos, I'd start with the first mode so Grandma still looks like Grandma.

Where Lightroom and Photoshop Still Fit

Aiarty isn't trying to replace your main editor, and I don't think it should. It reads raw files from most major cameras, and Aiarty explains how to enhance camera photos as part of a normal workflow. You can use it at either end of your edit. Run a noisy or soft file through Aiarty first, export it as a 16-bit PNG or a TIFF, and then do your color and retouching in Lightroom or Photoshop. You can also finish your edit as usual and use Aiarty as the last step to upscale before you send a file to the print lab. For a whole set of similar photos, like a dark reception or a stack of old scans, batch processing applies the same settings to every image in the folder.

Almost everything Aiarty does can technically be done in Photoshop if you know how. Rebuilding a soft face, cloning people out of a busy scene, or restoring a faded print takes careful masking and a lot of practice. On top of that, Photoshop is a subscription you pay for every month or every year without ever owning the software. Aiarty asks you to pick the model that matches your problem and move a slider. For novice retouchers, hobbyists, or working photographers who never learned advanced Photoshop, that's a much shorter learning curve.

How Aiarty Compares With Topaz on Price

The closest competitor to Aiarty is Topaz Labs, which makes popular upscaling and noise reduction software. Topaz is excellent software with a ton of features, including cloud rendering and a strong set of video tools. However, Topaz has moved entirely to subscriptions, and Adobe recently finished buying Topaz Labs. As of this writing, Topaz Photo costs $199 per year, or $169 for the first year with Topaz's own Black Friday discount. Gigapixel, its dedicated upscaler, is $149 per year.

Topaz Photo renews at $199 every year, while the Aiarty Image Enhancer lifetime license is a one-time $109 purchase. Prices from topazlabs.com and aiarty.com on November 11, 2026.

Aiarty's regular price for the Image Enhancer lifetime license is $109, and one license covers up to three computers. You pay once, the license never expires, and updates are included for life. If you only need to rescue a handful of photos each year, paying for software every single year adds up fast. Owning a tool that sits on your computer until you need it makes a lot more sense. If you've been curious about AI enhancement but didn't want another subscription, this is the more affordable way in. We first covered Aiarty Image Enhancer back in 2024 if you want to see where it started.

What I Liked

  • Everything runs locally, with no uploads, tokens, or credits.
  • The lifetime license includes updates and commercial use, and it covers up to three computers.
  • Each model has a clear job, so it's easy to pick a starting point.
  • One program handles noise, blur, upscaling, object removal, and old photo restoration.
  • It reads raw files and exports TIFF and 16-bit PNG files for further editing.
  • The free test images and settings guide make it easy to learn what each model does.

What Could Be Improved

The AI Eraser is still labeled experimental inside the app, although it handled my walkway example cleanly. The old photo examples also show that enlargement and restoration aren't the same as a full retouch. Dust specks on the playground print and the blue cast on the group portrait both came through. For a finished family print, plan on a few minutes with the color tools or a healing brush. More-Detail GAN v3 builds new texture that wasn't fully in the original file. For journalism, documentary work, or anything where accuracy matters more than polish, I'd stick with Real-Photo v3. The free trial also adds a watermark to exports and turns off batch processing, so you'll need the full version before you use it on client work.

Black Friday Special Offer and Final Thoughts

For $79, you own a tool that stays on your computer for good. The next time a relative hands you a box of faded prints or a client asks for a wall print from a cropped file, it will be ready.

The $109 lifetime license from the Topaz comparison above drops to $79 for Fstoppers readers during Aiarty's Black Friday sale this November. During checkout, choose the option for a coupon code, type FSTOPPERS into the box, and the price updates. The license covers three PCs, three Macs, or one PC and one Mac, and it comes with a free one-year license for both Aiarty Image Matting and VideoProc Converter. Aiarty also offers a 30-day money-back guarantee, and you can download the free trial of Aiarty Image Enhancer to run your own photos through each model before you buy. You can get the Fstoppers Black Friday deal here. This post is sponsored by Aiarty, but between the local processing and the one-time price, it's a deal I'd happily pass along to our readers anyway.

Images used with permission of Aiarty.

Read Entire Article