Image Tools

Remove Background is an image tool that recognizes the dominant foreground object in an image and removes its surroundings or background. The output is a clear .png image of the foreground object with either transparent or white background, that you can download and work with, or that will be automatically passed to other image processing tasks.

For example, if you need to upload product images from different sources, our AI can pick the highest-quality image, remove its background, and make it the main photo of the product, which will be visible in the product listing.

You can try how this service works in a public demo.

Background removal by AI helps businesses to automatically clear their product images and replace the original background with their own. It helps with consistency and visual merchandising in enormous collections with images from various sources.
It can be easily implemented into an automatic image processing Flow for your website. It is also a useful tool for graphic designers, web developers, photographers, and artists.

Our automatic background removal has its own public demo. If you want to use the service, log into our App. For production, you can connect to the service via API.

Remove Background is available within the Free plan. This plan with 3000 API credits allows you to process up to 100 images per month.

Image Upscaler is an image tool by Ximilar able to artificially enhance the resolution of images by up to 8x seamlessly and without the loss of details.

Ximilar provides several upscaling solutions. You can choose between 2x, 4x and 8x upscaling, and between Classic or GAN upscaling methods. It can be used multiple times on one image to achieve the highest resolution possible.

The Upscaler can be used both as a stand-alone tool or as a part of complex image-processing systems. As a tool, it helps graphic designers, web developers, photographers, and stock photo agencies.

Upscaler can also be a part of image-processing systems built with Flows. For example, you can use it to automatically upscale all low-resolution images (recognized by a recognition task), all user-generated content, or all images containing objects such as human faces or products that you sell (detected by object detection). This way, only high-quality images will be visible in product listings and on product pages.

The Classic Image Upscaler works exclusively with pixels. It multiplies each pixel in an image 2–8 times to achieve the desired resolution in a way that will not modify the image while adding detail to it. The resulting image is as true to the original image as possible. It is a reliable and consistent method for companies that need the contents of the image to stay precisely the same without any modifications and new elements added by artificial intelligence. It’s ideal for data such as CCTV footage, or images with delicate patterns and details.

GAN Image Upscaler (generative adversarial network), on the other hand, analyzes the colors, edges, corners, light and shade in the original image to add as relevant pixels as possible. It doesn’t only multiply, but generates new pixels, to provide natural-looking high-quality images. That is why the smallest details in the images can be slightly modified by this method. Because it helps with making the content more appealing, it is great for stock photos, as well as product images.

Image upscaling typically requires more GPU than other image-processing tasks, and therefore it can take a bit longer time. Upscaling can take from milliseconds to minutes, depending on the upscaling method you choose (classic or GAN), the size and the number of your images.

You can track the progress if you use the asynchronous API requests.

You can try the basic Image Upscaler as a stand-alone tool at StockPhotos.com.
To use the full potential and options of Image Upscaler, we recommend logging into our App or connecting via API. You can use the Upscaler both separately or combined with other services via Flows.

There are two basic ways of connecting to the API endpoints. A typical upscaling task uses synchronous requests (calls), meaning the user uploads an image, waits for it to be processed and is eventually provided with the result. This approach is typically used in public upscaling tools and is currently set for testing purposes in our App.
It can however be ineffective and time-consuming for companies that need to upscale large volumes of data at once and keep track of the progress. That is why we also provide an API endpoint for asynchronous requests. Asynchronous requests are a recommended way for using upscaling services.

The difference is that you can send multiple requests (specified by id), which will be queued and processed one by one, and you can send another (different) request anytime to track the progress of the job. We especially recommend this approach to companies that need to upscale whole databases, e.g. stock photo databases or e-shops with large product photo collections.

With asynchronous requests, you can also use Webhook, where we will send the information, and you will get a notification once the job is done.

Usage of our services is measured in API credits. You can see the full list o image-processing operations and their exact credit values in our Pricing. The amount of credits depends on the output resolution of the upscaled image.
If you are about to upscale large data volumes, let us know, and we will figure out a custom solution.