How ByteCut Works
ByteCut is a browser-based image preparation tool for people who need transparent cutouts, consistent square assets, and lighter WebP files without uploading images to a server. This documentation explains the processing flow, supported formats, export choices, AI cleanup behavior, privacy boundaries, and review checks before publishing an asset.
Quick summary
Runs in the browser
Source images are opened by your browser and processed on your device. ByteCut is not a cloud image library and does not store your original files.
Two main workflows
Use Remove Background when the background is the problem. Use WebP Optimizer when the image already looks right but needs a smaller delivery file.
Outputs are new files
ByteCut exports a separate WebP file. Your original image remains unchanged on your device.
Processing flow
1. Upload or drop an image
Start with a PNG, JPG, JPEG, or WebP file. Single-subject images with clear contrast are usually easier to clean than busy scenes.
2. Choose the workflow
Background cleanup creates transparent assets. WebP optimization resizes or compresses images that do not need transparency work.
3. Pick a practical preset
Food, Icon, Thumb, SNS, and Large presets help choose a useful size before export. Smaller presets are better for lists and UI; Large is better as a reusable source.
4. Review the preview
Use the checkerboard preview for transparent assets and check the final size before downloading. Look at the exact size where the image will be used.
5. Download the result
The exported WebP is a new local file. If the result looks wrong, adjust sensitivity, try the alternate cleanup path, or use a different source image.
Supported formats and outputs
ByteCut accepts PNG, JPG, JPEG, and WebP files in the editor. Background cleanup is designed for transparent WebP exports. Optimization produces resized or compressed WebP copies for websites, apps, Open Graph images, menus, thumbnails, and asset libraries.
PNG input
Best for source graphics, screenshots, and images that may already include transparency.
JPG/JPEG input
Useful for photos and generated images, but compression artifacts can make edge cleanup harder.
WebP input
Good for existing web assets that need resizing, square fitting, or another optimized export.
WebP output
Supports transparency and usually keeps file size lower than comparable PNG exports.
Choosing the right mode
Remove Background
Use this for stickers, food illustrations, characters, logos, and icons where the subject should sit on a transparent background.
WebP Optimizer
Use this for photos, thumbnails, social images, hero images, or generated visuals that already have the right background.
Resize Icons
Use square presets when a set of assets needs consistent dimensions in a menu, grid, app UI, or design system.
Export settings and presets
Food / Thumb presets
Use compact sizes for menu thumbnails, product cards, and repeated UI lists where fast loading matters.
Icon preset
Use icon-sized output for app UI, category graphics, sticker trays, and small reusable visual elements.
Large preset
Use Large when you need a cleaner source for later editing or a high-density display. Avoid it for every thumbnail because files get heavier.
Quality setting
Quality 80 is a practical default for many WebP exports. Raise it when small text, line art, or detailed food textures look too soft.
AI cleanup and fallback behavior
Some images need more than the fast bright-background cleanup. When AI cleanup is enabled, ByteCut may download a model on first use and run it locally in the browser. The model path can be slower and may use more memory, so ByteCut keeps the quick cleanup path available for simple backgrounds or when AI output looks unreliable.
When AI helps
Try AI for hair, character props, handles, interior holes, or backgrounds that are not plain white.
When quick cleanup is safer
Food stickers, white-background graphics, and simple illustrations can look cleaner with the fast path.
What to do if AI fails
Retry with fast cleanup, use a smaller source image, or choose a source with clearer contrast between subject and background.
Privacy and local processing
ByteCut is designed so image processing happens in the browser. Source files, previews, masks, and downloads are handled locally during the session. Contact form messages are separate from image processing and are used only so ByteCut can reply to a support request.
No image upload for editing
The editor does not send your source image to a ByteCut server for cleanup or optimization.
No account library
ByteCut does not create a hosted gallery, project folder, or cloud history of processed images.
Third-party scripts
Analytics or advertising scripts may load on the site, but they do not receive the image files you process in the editor.
Performance expectations
Large images use more memory
Very large photos can slow the browser or fail on low-memory devices. Resize the source first if the browser becomes unstable.
AI can take longer on first use
The first AI cleanup can be slower because model files may need to download and initialize.
WebP exports are for delivery
Keep original source files separately. Use ByteCut exports as smaller files for publishing, not as your only archive copy.
Limitations and manual review
Low-contrast edges
White plates on white backgrounds, pale clothing, glass, and soft shadows can blend into the background.
Fine detail
Hair, thin props, small handles, and transparent interiors should be checked on the checkerboard preview.
Generated food images
AI-made food can contain impossible utensils, repeated textures, fake text, or inaccurate ingredients. Review before publishing.
Brand or legal review
ByteCut prepares files, but you are responsible for rights, brand fit, platform rules, and whether the image accurately represents your product.
Troubleshooting checklist
Background remains visible
Increase cleanup sensitivity, try AI retry, or use a source image with clearer contrast.
Subject is damaged
Lower the cleanup aggressiveness, use quick cleanup instead of AI for simple graphics, or keep more padding around the subject.
Output looks blurry
Avoid upscaling small images. Export closer to the original size or start from a higher-resolution source.
File is larger than expected
Try a smaller preset or lower WebP quality. Transparent graphics exported too large can still become heavy.