Enlarge without falling apart
Enlarge up to 5x while the model reduces noise and reconstructs edges rather than smoothing them away. Product photos and portraits hold up at sizes the original file was never meant to reach.
Topaz Upscale enlarges images up to 5x while reducing noise, correcting blur, and restoring detail. It is built on Topaz Labs' image models, with separate model families for enhancement, sharpening, and denoising rather than one setting that tries to do everything.
In Phygital+ it sits at the end of a chain: generate or upload an image, then push the frame you actually want through the upscaler. Face recovery, subject detection, compression repair, and film grain are all exposed as controls on the same node.
Product photos for e-commerce, prints at larger sizes, portraits, and archive material
Photo, e-commerce, and print teams preparing finished images for a bigger format
Faithful enlargement with per-problem models, face recovery, and repair controls.
Enlarge up to 5x while the model reduces noise and reconstructs edges rather than smoothing them away. Product photos and portraits hold up at sizes the original file was never meant to reach.
Separate model families for enhancement, sharpening, and denoising, each with variants for the specific problem: lens blur, motion blur, missed focus, high-ISO grain, CGI, low-resolution sources, images containing text.
Enable face recovery on portraits and set how strongly it applies. Creativity decides whether the recovery stays faithful to the original face or interprets it, so identity survives the upscale on people-heavy images.
Subject Detection limits the effect to the foreground, the background, or the whole frame. Useful when the subject is clean but the background is noisy, or the other way around.
Compression artefacts, blocking, and ringing have their own controls, and the low-resolution model targets weak sources. Old assets and heavily compressed files get a second life instead of a reshoot.
Topaz Upscale fits at the end of a workflow, where a finished image has to survive a bigger format than it was generated or shot at.
Connect an image, choose a model and multiplier, and run.
Connect an image to the node. It can come from a generation node on the same canvas or from an upload, at up to 4096 px per side. The cleaner the source, the better the result: upscaling restores structure, it does not invent what was never captured.
Start with Enhance Standard V2 and the smallest multiplier that meets your output size. Switch model family when the problem is specific: Sharpen for blur or missed focus, Denoise for grain, Enhance Low Resolution for weak sources, Enhance Text Refine when the image contains type.
For portraits, enable Face Enhancement and keep creativity low so the identity survives. Use Subject Detection to limit the effect to foreground or background. Add a small amount of film grain at the end if the output looks too smooth, then export.
One subscription across 30+ AI models, with no per-tool credit balances or separate signups. Credit cost per run is shown live in the node before you start it.
500 weekly credits to test the node and see the output quality. No credit card required.
45,000 monthly credits with cheaper per-credit pricing, commercial rights, and full downloads.
90,000 monthly credits, up to 10 seats, shared workspace, centralized billing, and priority support.
210,000 monthly credits, API and integration support, unlimited seats, and dedicated management.
Full specs for Topaz Upscale in Phygital+.
How Topaz Upscale compares with the other enhancement nodes in Phygital+.
Put upscaling at the end of a generation chain instead of running it as a separate tool.
Generate at the model's native size, then upscale the frame you picked instead of paying render time on every variant.
Upscale first, then cut the subject out. A cleaner edge on a larger image means a cleaner mask downstream.
Use the upscaled frame as the first frame for a video node so the clip starts from the sharpest version of the shot.
Send an upscaled still into a vectoriser when the asset needs to scale without any resolution ceiling at all.
Everything you need to know about Topaz Upscale in Phygital+.
Topaz Upscale is the Phygital+ node built on Topaz Labs' image models. It enlarges images while reducing noise, correcting blur and compression damage, and restoring detail. It is a restoration tool first: the goal is a bigger version of the image you already have, not a reinterpretation of it.
From 1x to 5x. Use the smallest multiplier that meets your output size, because larger factors give the model more room to introduce artefacts. Input images are accepted up to 4096 px per side.
The model list covers three families. Enhance models handle general enlargement, low-resolution sources, CGI, and text refinement. Sharpen models target specific problems: lens blur, motion blur, missed focus, portraits, wildlife. Denoise models handle grain and high-ISO noise. Enhance Standard V2 is the default and the right starting point for most images.
Turn on Face Enhancement when people are the subject. Enhancement Strength controls how much facial detail is recovered, and Enhancement Creativity decides whether recovery stays faithful or interprets. Start creativity near zero for restoration work and raise it only if the face still looks soft.
Magnific is generative: it adds detail and texture, which is what you want for AI images and design assets that should look richer. Topaz is restorative: it recovers what is already in the file, which is what you want for photographs, product shots, and anything where accuracy matters more than richness.
Yes, but within limits. Fix Compression and Decompression Strength target JPEG blocking and ringing, and the low-resolution model is built for weak sources. Upscaling cannot recover detail that is not in the file, so a clean source always beats a heavily damaged one.
Add it after upscaling. AI enhancement can leave an image looking too smooth and synthetic, and a small amount of grain puts the texture back. Keep the size small and the strength low; heavy grain reads as a stylistic choice rather than a finishing touch.
Usage is credit-based and included in every plan, from Free up to Enterprise. The credit cost per run is shown in the node before you start it, so there is no separate Topaz licence or subscription to buy.
Outputs created on any paid plan come with commercial-use rights, including images you upscaled from your own photography.
Browse the full catalog of 30+ AI models available in Phygital+.
Join 100+ teams using Phygital+ – every model in one workspace
Try Phygital+ free