Image Slicer

Slice a single image online: split by a fixed grid (by columns/rows or cell size), auto-detect scattered independent regions by transparency or background color, and manually draw, move and resize each slice box on the canvas with pixel-precise control. Batch-crop and pack all slices into a ZIP in one click, with PNG / JPG / WebP output and optional transparent-edge trimming. Everything runs locally in your browser with no uploads — secure and fast.

1Upload image

Features

  • Fixed-grid slicing: split by columns × rows, or derive from cell width/height, with offset and row/column spacing
  • Auto-detect scattered regions: connected-component detection by alpha channel or background color to box each independent element
  • Auto-detect with adjustable tolerance, minimum size, outward padding and neighbor merging for cleaner, more accurate boxes
  • Manual box editing: drag to create slices, resize with eight handles, move as a whole, or enter exact coordinates for pixel-precise placement
  • Mix all methods: generate a grid / auto-detect first, then fine-tune, add or delete boxes manually
  • Batch export: crop all slices at once and pack them into a ZIP — PNG / JPG / WebP, with optional transparent-edge trimming
  • Runs entirely locally in your browser — files are never uploaded

FAQ

Are my images uploaded to a server?

No. Image decoding, region detection, slicing and packing all happen locally in your browser — files are never uploaded to any server, secure and fast.

How does "auto-detect scattered regions" work?

The tool treats transparent pixels (or pixels close to the background color) as background, then uses a connected-component algorithm to find disjoint independent elements and takes their bounding boxes as slices. It works well for background-free sprite sheets, icon sets and loose images. If elements stick together, lower the tolerance or reduce the merge gap; if one element is split into several boxes, increase the merge gap.

Can I edit the grid or detected results manually?

Yes. After generation, every box can be dragged, resized with handles, or edited via precise X / Y / width / height inputs below; you can also drag to create new boxes or delete any of them.

How are the exported files named?

Boxes are numbered in reading order (top-to-bottom, left-to-right) as "prefix_001.png", with a customizable prefix, and all slices are packed into a single ZIP for download.