Convert and Compress
PNG, JPEG, WebP and AVIF, with a quality control for the lossy formats. Compression keeps the source format and reports the real change in bytes, including when there is none.
Image Toolkit
Convert, compress, resize, crop and remove backgrounds
Ten image jobs in one page, all of them on your own machine
Web Tool
Official product page
Publisher:
POLPROG
Official product page
10
Panels
4
Output formats
7
Direct routes
0
Bytes uploaded
01 / Image Toolkit
About
Image Toolkit puts ten jobs behind one file intake: convert, compress, resize, crop, rotate and flip, watermark, remove background, metadata, social sizes and favicon. You drop the files once and switch panels; the queue does not have to be rebuilt for every operation.
Everything runs on a canvas in the tab. No image is uploaded, and there is nothing on a server to delete afterwards. That is also why every number the page shows - the saved bytes, the output dimensions - appears only after a batch has actually run.
Background removal is a real model, not a colour key. The panel downloads a 4.4 MB self-hosted U-2-Net segmentation model and runs it through ONNX Runtime in WebAssembly, once, the first time the panel is opened. It produces a saliency mask rather than an alpha matte, so hair and fur come out with a harder edge than a paid matting service would give. The page says so, instead of leaving somebody to find out on a portrait.
Metadata stripping is lossless. Removing EXIF, GPS and colour profiles rewrites the container and leaves the compressed image data untouched, so a photograph does not quietly lose a generation of quality to have its location removed. Compression, by contrast, re-encodes - and a file that could not be made smaller is reported as already optimal rather than as a fabricated saving.
02 / Use case
What it solves
The problem
- Converting one image to WebP means uploading a photograph to a site you have never heard of.
- Crop, resize, favicon and background removal are four different tools with four different queues.
- Stripping the GPS tag out of a photo usually re-encodes it and costs a generation of quality.
The outcome
- Ten operations behind one intake, on your own machine, with nothing uploaded.
- PNG, JPEG, WebP and AVIF in and out, in batches of up to thirty files.
- Metadata removal that rewrites the container and leaves the pixels byte for byte identical.
Who it is for
- Anyone who needs a smaller image and does not want to hand it to a stranger.
- Developers and designers producing favicons, OG cards and platform-sized exports.
- People sharing photographs who would rather not share where they were taken.
03 / Key Features
Key Features
The most important capabilities, explained in practical terms.
Resize and Crop
Resize by width, height or percentage with an optional aspect-ratio lock, and crop with a draggable frame, fixed ratios or exact pixel entry. Neither ever enlarges past the source unless you ask.
Rotate and Flip
Ninety-degree rotation in both directions and horizontal or vertical flips, applied to the whole batch at once.
Background Removal
A self-hosted U-2-Net model runs in WebAssembly on your device. The 6.1 MB download - the ONNX runtime and the 4.4 MB weights together - happens once, when the panel is first opened, and never on any other page.
Metadata, Removed Losslessly
EXIF, GPS coordinates, camera identifiers and colour profiles are stripped by rewriting the container. The compressed image data is not touched, so nothing is re-encoded.
Social and Favicon Sizes
Open Graph, Twitter, LinkedIn, Instagram and YouTube dimensions from one source image, and a favicon set with the sizes browsers and platforms actually request.
Seven Direct Routes
Compress, resize, crop, remove background, remove metadata, favicon and social resizing each have their own page, so a search for one of them lands on that panel rather than on a tab strip.
Watermark
Text over every picture in one pass, in nine positions or repeated across the frame. The size is a share of the image width, so one setting suits a batch of differently sized photos.
Compress to a size, not to a number
Ask for 200 KB and the encoder is run repeatedly to find the highest quality that fits, shrinking the picture only when quality alone cannot get there. A before and after slider sits beside it, and the size under it is the size of the file the button would write.
What a canvas cannot keep
An animated GIF, APNG or WebP is flagged before any work starts, because a canvas holds one frame. So is an embedded colour profile, because the output is sRGB. Both are said out loud rather than discovered after the download.
04 / Download
Product access
A clear path from the official source to your first successful workflow.
Choose the right platform
Use one of the official links available for your device or browser.
Install or open
Follow the store instructions or open the web application. No third-party installers.
Configure and begin
Review the options, choose your preferences and start with the core workflow.
Official access options
05 / Technology
Technology behind the product
A transparent look at the core technologies used to build and maintain this product.
JavaScript
06 / Security
Privacy & Security
Your data stays on your device. Always.