Split an image into equal parts, on a grid

Splitting an image into equal parts

Some tasks call for exactly the opposite of cropping an area: instead of keeping just one part of the image, the goal is to divide the whole thing into regular pieces, without losing a single pixel along the way. A wide, panoramic image posted as a sequence in a carousel — each slide being a vertical slice of the original panel. A scanned sheet with several items arranged in a grid — receipts, cards, labels — that need to become separate files. A sheet of cut-out pieces, or a large panel that needs to be printed on smaller sheets and reassembled later. In every one of these cases, working out by hand where each cut should fall is tedious and easy to get wrong by a pixel.

This tool does that split precisely: you set how many columns and rows you want, and it works out each cell's boundaries from the image's actual dimensions — nothing to install, and the file never leaves your computer, because all the processing runs in your browser.

A configurable grid, with shortcuts for the common cases

The split accepts 1 to 10 columns and 1 to 10 rows, combined however you like — 2 columns by 3 rows makes 6 parts, for example. For the most common cases, there are ready-made shortcuts: 2 parts horizontally or vertically, 3 parts horizontally or vertically, 4 parts in a 2×2 grid, and 9 parts in a 3×3 grid. A preview shows the cut lines overlaid on your uploaded image, with each cell numbered in the order it will be exported — useful for checking the split before generating any file, especially when publishing order matters.

No pixel gets left out

When the image's width or height isn't an exact multiple of the number of columns or rows you chose, there's a remainder left over — a 1000-pixel-wide image split into 3 columns, for instance, doesn't divide evenly into three exact 333-pixel parts. Instead of just cutting that remainder off, or silently squeezing the last part differently from the others, the tool spreads those extra pixels across the first cells of the grid, which end up one pixel larger than the rest. The difference isn't visible to the eye, but it guarantees the whole image is covered by the generated parts, with no lost strips at the edges.

Predictable export order and file names

Each part is numbered in standard reading order: left to right, top to bottom. The first cell (top-left corner) becomes "part-01", the next one in the same row "part-02", and so on through the last cell, in the bottom-right corner. The numbers come zero-padded, so the files stay in the right order even when opened in a program that sorts names as plain text. That predictability matters especially for anyone publishing the parts in sequence, like in a carousel: the order the files are generated in is already the right publishing order, with nothing to rearrange afterward.

Fixed ratio: when a bit of extra cropping makes sense

By default, each part comes out in whatever natural shape the grid division produces — a cell from a rectangular image comes out rectangular too. But sometimes the final destination calls for a specific ratio, like square (1:1) for a social feed, or 9:16 for a story. Turning on a fixed ratio crops each cell centered until it matches that ratio, discarding the excess at the cell's edges — an extra crop, on top of the grid split. This mode is off by default because it discards content; the interface flags it whenever it's active.

Batch mode and output format

Upload more than one image, and the same grid is applied to all of them — but each image is split based on its own dimensions, so a larger image produces larger parts, keeping the same proportional split. In the batch result, the files from each source image are organized into their own subfolder inside the downloaded ZIP, instead of all mixed into one folder. Output can be JPG, PNG, or WebP, with a quality control on the first two.

What this tool doesn't do

It doesn't analyze the image's content to decide where to split — the grid is always regular, based only on the number of columns and rows you chose, not on what appears in each region. It doesn't reassemble the parts back into one image afterward; once split, putting them back together is a separate task. And it doesn't upscale or improve the original content's quality: each part comes out at the same quality and resolution that region already had in the source image, never sharper than it was.