Upscayl AI
Open-source Desktop app · Cloud workflow available

Free, Open-Source AI Image Upscaler for Desktop and Cloud

Enlarge low-resolution photos, artwork and product images with a practical AI image upscaler. Use Upscayl Desktop for compatible local processing, or choose Cloud for a browser-based workflow.

Try Upscayl Online Download Desktop
Compare Desktop and Cloud
  • Windows, macOS and Linux
  • Local Desktop processing
  • Multiple model options

Upscaling can improve the appearance of suitable edges and textures, but it cannot recover every detail that was never captured. Test one image before processing a full batch.

Interactive image preview Drag to review the result
Photo example
Pink vintage car in a desert scene after a restrained image enhancement
Pink vintage car in a desert scene before image enhancement
Before After
Review edges and texture Tip inspect at final size
See the difference on real images

Judge an image upscaler by the details that matter.

Review the same image at its intended output size. Look for natural edges, believable texture, accurate shapes and artifacts that could distract from the final use.

Three hikers in a snowy mountain setting used to review faces, fabric and landscape detail
Photo detail

Faces, fabric and fine contrast

Check facial identity, clothing texture, snow edges and distant scenery without accepting halos or painted-looking detail.

Compare photo examples
Elephant in warm grassland light used to assess wildlife texture and edge quality
Natural texture

Wildlife, foliage and difficult edges

Inspect fur or skin texture, grass patterns and soft atmospheric areas for repeating detail or excessive sharpening.

Vanilla ice cream product photograph used to inspect product edges and surface texture
Product images

Clean silhouettes and accurate materials

Review packaging or product boundaries, reflections and small textures without inventing marks, labels or materials.

Upscale ecommerce images
Free Desktop app Open-source project Local processing option Cloud and API workflows Built for real image tasks
One product family, three paths

Choose where your images should be processed.

Start with the outcome and the device you have. The site should make this decision clear before asking you to compare every feature.

Desktop

Free, open-source and local.

Install Upscayl on a compatible Windows, macOS or Linux computer and keep supported processing on your device.

  • Local image workflow
  • Batch processing
  • Multiple and custom models
Download for Desktop
Cloud

Upscale from a browser.

Use a browser-based workflow when local hardware is not a good fit or when you want access across different devices.

  • No Desktop installation
  • Additional model choices
  • Image editing workflow
Open Upscayl Cloud
API

Build repeatable image processing.

Connect an image-upscaling task to a product, content pipeline or internal workflow using the documented API.

  • Structured task requests
  • Multiple supported models
  • Developer documentation
Explore the API
Not sure which one fits? Compare Desktop and Cloud
Tools for different source images

Useful controls without a complicated first step.

Begin with a sensible model, inspect the output, and adjust only when your source image needs a different approach.

Batch workflow

Prepare a set of images together.

Apply a consistent workflow to a folder of product images, scanned photos or creative exports instead of repeating the same steps one file at a time.

Explore batch upscaling
Privacy choice

Process locally with Desktop.

A local path is useful for compatible hardware and images you do not want to send through a browser workflow.

Understand local processing
Model choice

Match the model to the image.

Photos, compressed graphics and illustrations can need different treatment. Use examples rather than guessing from model names.

Compare Upscayl models
Face enhancement

Give portraits a focused pass.

Use face enhancement selectively for suitable portraits and old photos, then compare the result at the size you actually need.

Explore face enhancement
Larger output

Build enough resolution for the next step.

Prepare a larger working file for print, design layouts, product pages or high-density displays—while keeping expectations tied to the source.

Plan a high-resolution output
A more useful model guide

Start with the image type—not the most dramatic model name.

Start with a representative image type, then compare genuine first-party tests before choosing a model for a full image or batch.

Suggested starting point

Natural-detail model

Use a model that protects believable texture and avoids adding an overly processed look to skin, foliage and fine contrast.

See model tests and limitations
Outdoor photography example with natural skin, fabric, foliage and landscape detail
Source review Texture · edges · compression
A second look, at the detail that matters

Review the image before you commit the whole batch.

Use the divider to check edges, texture and contrast at the part of the image that matters to the final job. The example is intentionally interactive so visitors understand the task before choosing a workflow.

Start with a representative crop. Fine detail is easier to judge when the source stays visible.

Move the divider yourself. A real comparison builds more confidence than a claim alone.

Explore model guidance
AI-art detail preview

Check the output at working size.

Drag to inspect
Clear AI artwork showing a person with robotic hands after image enhancement
Same AI artwork before enhancement with softened facial, clothing and mechanical detail
Source detail Reviewed output
Focus Face, fabric and mechanical edges Workflow Drag to review before export
Start with a real task

Built around where the image needs to go next.

Explore all solutions
01

Photography

Prepare selected images for larger displays, crops and client deliverables.

 02

Old photos

Build a careful scan, enhancement and review workflow for family images.

 03

Printing

Connect pixel dimensions, print size and viewing distance before exporting.

 04

E-commerce

Create a consistent image set for product pages and marketplace listings.

 05

AI art

Enlarge digital artwork while checking lines, texture and generated detail.

 06

Design teams

Choose Desktop, Cloud or API processing based on volume and collaboration.
Evidence before adjectives

Show how the result was made.

A trustworthy upscaler site should name the version, model, scale, processing environment and source image. It should also show weak outputs—not only the cleanest example.

01

Same source image used across model and product comparisons.

02

Visible test settings with dates and relevant hardware details.

03

Honest limitations when texture, faces or edges do not improve cleanly.

Read testing methodology
Transparent test record Transparent setup
Transparent image upscaling test record comparing source, improved result and a limited result with visible artifacts
Input
Same source crop
Scale
Visible per test
Model
Named with version
Reviewed
Clean and weak outputs
A clear first run

From source image to reviewed output.

The interface should make the next decision obvious without hiding quality controls from users who need them.

  1. 1

    Choose a path

    Download Desktop for a compatible local workflow or open Cloud in your browser.

  2. 2

    Add the source

    Use the best available original and avoid repeatedly compressed copies when possible.

  3. 3

    Select a model

    Start with the image type, then compare a small crop before committing to a batch.

  4. 4

    Review at final size

    Check important faces, edges and textures at the display or print size you need.

For product and content teams

Move repeatable image upscaling into your workflow.

Use the API when image processing needs to connect with an upload flow, catalog pipeline or internal tool. Keep the marketing page concise and link developers to the current technical reference.

Explore the API Read API docs
start-task.js
const form = new FormData();
form.append("file", imageFile);
form.append("model", selectedModel);

const response = await fetch(
  "/api/upscale-task",
  { method: "POST", body: form }
);
Learn before you process everything

Guides for better source images and cleaner decisions.

Browse all guides
Common questions

What to know before your first upscale.

These answers are deliberately practical. Product capabilities and requirements should be reviewed whenever the software changes.

Visit the support center

Yes. Upscayl Desktop is presented as a free, open-source image upscaler. Review the current download and license information before distributing or using it in a commercial workflow.

Compatible Desktop workflows can process images locally. Cloud and API workflows require an internet connection because images are processed through an online service.

Upscayl Desktop is available for Windows, macOS and Linux. Check the download page and documentation for current versions and hardware requirements.

Choose Desktop when local processing and an installed application fit your workflow. Choose Cloud when you want browser access or do not have suitable local hardware.

The useful choice depends on the source image. Photos, compressed images and illustrations can respond differently, so compare representative crops before processing a large batch.

Upscaling can provide more pixels for a print workflow, but final quality also depends on the original image, print size, viewing distance, sharpening and the printer.