Skip to main content
MindStudio
Pricing
Blog About
My Workspace
Image Editing

Upscale Image

The Upscale Image block increases the resolution of an existing image using AI upscaling, taking an image URL as input and returning a higher-resolution version hosted on a CDN.

Increase image resolution using AI upscaling

The Upscale Image block increases the resolution of an existing image using AI upscaling, taking an image URL as input and returning a higher-resolution version hosted on a CDN. It supports three target output resolutions — 2K, 4K, and 8K — and two engine tiers: standard and pro. The upscaled image is always returned as a PNG file accessible via a CDN URL.

The block requires two core parameters: the URL of the source image and the target resolution. You also select which upscaling engine to use, and optionally specify a destination variable where the output URL will be stored for use in later steps of your workflow.

This block fits into workflows where image quality needs to be improved before display, download, or further processing — such as e-commerce product image pipelines, content generation workflows that produce images at lower resolutions, or any workflow where user-submitted images need to meet a minimum resolution threshold before being used.

What you can build

Real-world workflows powered by the Upscale Image block.

E-Commerce Product Images

Automatically upscale product photos submitted by vendors to meet minimum resolution requirements before publishing to a storefront.

AI Image Generation Pipeline

After generating images with a text-to-image block, pass the output through the Upscale Image block to produce a higher-resolution final asset.

Print-Ready Asset Preparation

Upscale digital images to 4K or 8K before sending them to a print fulfillment service that requires high-resolution files.

User Photo Enhancement

Accept low-resolution user-uploaded photos and return an upscaled version for display in a profile, portfolio, or gallery application.

Real Estate Listing Photos

Enhance property images uploaded by agents to a higher resolution before they are published to a listing page.

Archival Image Restoration

Process scanned or digitized archival images through the upscaler to produce higher-resolution versions suitable for digital preservation or publication.

Ready to add Upscale Image to your workflow?

Get Started Free

Common questions about Upscale Image

What parameters are required to use the Upscale Image block?

The block requires two parameters: imageUrl, which is the URL of the source image to upscale, and targetResolution, which must be one of '2k', '4k', or '8k'. You also need to select an engine tier — either 'standard' or 'pro'. The destinationVar field is optional and specifies which workflow variable will store the output URL.

What does the block return?

The block returns an imageUrl field containing a CDN-hosted URL pointing to the upscaled image. The output file is always in PNG format, regardless of the format of the source image.

What is the difference between the standard and pro engine options?

The interface defines two engine tiers: 'standard' and 'pro'. These represent different upscaling engine quality levels. The metadata does not specify further technical differences between them beyond the quality tier designation.

What kinds of workflows commonly use this block?

This block is commonly used in image generation pipelines where output resolution needs to be increased, in content or media workflows that require images to meet a minimum resolution before publishing, and in user-facing applications where uploaded images need to be enhanced before display or download.

How do I use the upscaled image in later workflow steps?

You can specify a variable name in the destinationVar field. Once the block runs, the CDN URL of the upscaled PNG will be stored in that variable, making it available for reference in subsequent blocks within the same workflow.

Add Upscale Image to your workflow

Build powerful AI workflows with drag-and-drop blocks. No coding required.