Product changes

We updated our Service Level Agreement (SLA) and added Maintenance Window. The agreement now includes the "Scheduled Maintenance" term.

This means we will not include possible downtimes within scheduled maintenance in our Monthly Uptime Percentage (MUP) calculations.

We will inform you about scheduled maintenance and possible consequential downtimes. In every message regarding scheduled maintenance, we will provide a description of what operations we will perform and which systems may be affected.

Indeed, we strive to minimize possible downtimes at all times.

We added Photo Filters 📷 You can now pick one of the 40 filters that best suits your image and apply it on-the-fly with our Image Transformations.

Original Image

Filtered, gavin

Filtered, iorill

Applying a filter goes like this:

-/filter/:name/

or

-/filter/:name/:amount/

Where :name stands for one of the 40 elven names we chose for our filters, and :amount controls the strength of each of them.

You can discover filter names and possible :amount ranges in our docs.

You can now adjust image colors on-the-fly with our Image Transformations.

/vibrance/-50/

Original image.

/vibrance/50/

We added seven new transformations based on Look-Up Tables (LUTs). LUTs provide a precise way of taking specific RGB image values and modifying them to other ones. LUT color adjustments are fast too, not just precise. Color adjustments share the same syntax, e.g.:

-/brightness/:value/

Here's the list of color properties you can control:

• Brightness
• Exposure
• Gamma
• Contrast
• Saturation
• Vibrance
• Warmth

You can find thorough descriptions of the transformations along with their :value ranges in our docs.

We updated our Image Transformations with a new operation, Unsharp Masking.

Original image. 16Kb

/blur/100/-50/ 19Kb

/blur/100/-120/ 22Kb

Unsharp Masking can be used for enhancing images by improving their sharpness. The effect is also called "Local Contrast Enhancement." Here's how the syntax goes:

-/blur/:strength/:amount/

Technically, we did not introduce a separate operation for Unsharp Masking. Instead, we added the :amount parameter to the blur transformation. Negative :amount combined with reasonably high blur :strength provides the Unsharp Masking behavior.

See how improving local contrast looks like in our docs or read more about Unsharp Masking in this Wikipedia article.

We increased the performance of the sharp Image Processing operation. Image sharpening is one of the most commonly used on-the-fly operations on Uploadcare.

Our internal benchmarks show that we increased the operation performance by up to 5x. This means that the sharpening-enabled image operations will become faster when output images have not been cached by our CDN yet.

No actions are required on your end.

You can now get larger file groups as ZIP or TAR archives. A new limit of a total uncompressed size of files in a group you archive is 2 GB.

Check out our docs for details on file groups and archiving.

Introducing Video Processing. You can now optimize videos, transcode them, cut fragments, and generate thumbnails.

Video Processing is available to all our customers and works via REST API, here is how:

• You post a processing job by making a REST API request; all the processing params are passed in the request.
• Once the job is finished, you get a UUID of a processed output.
• At the same time, the output becomes addressable via a CDN URL.

Here is what you can do with videos:

• Resize
• Adjust quality
• Transcode
• Cut fragments
• Generate thumbnails

Check out the complete Video Processing documentation here or drop us a line in case you have any questions.

Before the update, producing higher quality JPEGs from their low-quality originals could lead to increased output file sizes.

Now, if no destructive operations are applied, your output JPEG quality is limited to the initial input quality. I.e., when you provide low-quality JPEG inputs, using higher quality presets will neither increase file sizes nor change your initial compression. Learn more.

You can now convert GIFs to either H.264 or WebM video formats on the fly via our new gif2video operation. Video files are generally much smaller than GIFs hence their delivery is significantly faster.

The gif2video feature is disabled by default and should be explicitly enabled for your project. Learn more in our docs.

Now, instead of iterating through individual files and downloading them one by one you can get a whole pack in one go. We've introduced the archive operation for that. It allows you to get groups as either zip or tar archives.

Current limits are:

• total uncompressed size of files in a group should not exceed 100 MB
• archives include original files only — CDN API operations info is discarded

From this point on you can detect faces in photos with Uploadcare. We're happy to announce the detect_faces CDN API operation.

The operation gives you coordinates of faces found in your images in a neat JSON:

{"faces": [
  [120, 380, 560, 560],
  [1235, 310, 635, 635],
  [2375, 110, 740, 740]
]}

Use it to further improve your image workflow or build stuff from scratch!

You can now upload WebP images with our widget, process those on-the-fly using Uploadcare CDN or convert your existing images to WebP by adjusting their format.

You are now able to extract colors from your images with the main_colors CDN API operation. Basically, the operation is about analyzing an image in order to extract a given number of dominant colors as a list of RGB values.

We've increased CDN image dimensions limits. This means that you can retrieve larger images from CDN and process larger original files from your users.

Maximum source image resolution is increased from 50 to 75 Megapixels. This allows, for example, to handle panoramas from iPhone 6s.

Maximum output resolution is increased from 2048×2048 to 3000×3000 pixels (look at this example).

Default values for -/preview/ operation stay the same, so preview without arguments will produce images with resolution up to 2048×2048 to save bandwidth.

We've added an auto enhance operation to our CDN image processing API. It automatically tunes levels, contrast and sharpening.

So instead of original on the left you get enhanced version on the right:

The feature is available on all plans with no extra charge :)