Image Edit Methodology

Image Edit reads supported files on your device using browser image APIs. It accepts PNG, JPEG and WebP, with maximum inputs of 16 MiB, 24 megapixels and 16,384 pixels on either side. Generated outputs are limited to 16 megapixels and 8,192 pixels per side. Input limits reduce memory demands but cannot guarantee that every permitted image will decode on every device.

Geometry and edits

Locked resizing computes the missing side from the source aspect ratio and rounds to a whole pixel. Crops use integer coordinates measured from the top left and must remain inside the source. Rotations use 0°, 90°, 180° or 270° clockwise; flips apply before rotation. Quarter-turn rotations swap width and height. Brightness, contrast, saturation and grayscale adjust RGB values; the alpha channel is retained in PNG output.

Encoding and size

Edits are drawn to canvas and exported as a new still image. PNG and WebP can retain transparency; JPEG uses white for transparent areas. Export support depends on the browser. A quality setting is an encoder input, not a universal quality percentage. File-size saving is (original bytes − output bytes) ÷ original bytes × 100, and a negative result means a larger output.

Metadata and Base64

Canvas exports do not copy the original file’s metadata fields. They are not a tool for removing private information visible in pixels. Base64 encodes the original bytes unchanged, retaining metadata; its encoded length is 4 × ceil(bytes ÷ 3), excluding the data URL prefix. Animated images are outside the editing workflow and canvas exports are still images. The tools do not promise color-managed print proofs, AI restoration, background removal or universal file support.

For worked examples, see the image guides. Review previews and exported files before sharing.