Image Grid Splitter
Split one local JPG, PNG, or WebP into an exact grid of up to 100 full-resolution PNG tiles without uploading the image.
Privacy: your files never leave your device. All processing happens locally in your browser.
How to use
- 1.Choose a local JPG, PNG, or WebP within the displayed file and pixel limits.
- 2.Enter 1–10 rows and 1–10 columns, then select Split image.
- 3.Check each numbered tile's dimensions and download the PNG files you need.
About Image Grid Splitter
Image Grid Splitter divides one local image into a rectangular set of separate PNG files. Choose a JPG, PNG, or WebP, enter the number of rows and columns, and select Split image. The result lists every tile in top-to-bottom, left-to-right order with its exact dimensions and an individual download link. File decoding, pixel extraction, PNG encoding, and download creation all happen inside the current browser tab; the source is never uploaded to Lizely.
The tool accepts from one to ten rows and one to ten columns, for a maximum of 100 tiles. It uses the image's natural pixel dimensions rather than the scaled preview size. A 1200 by 900 image split into three rows and four columns produces twelve 300 by 300 tiles. A size that is not evenly divisible is handled deliberately instead of being rounded to a misleading common size. Boundaries are calculated from proportional integer coordinates, so adjacent tiles meet exactly and every source pixel belongs to one tile.
For example, dividing a 10 pixel width into four columns creates widths of 2, 3, 2, and 3 pixels. The difference between the widest and narrowest tile is at most one pixel, the four widths sum to ten, and no narrow strip is discarded. The same rule is applied independently to the height. The tool also rejects a grid that would create a zero-pixel tile, such as three rows from an image only two pixels high.
Each tile is rendered on a natural-size offscreen canvas. The browser copies only that tile's source rectangle, without stretching, interpolation, filters, borders, padding, or recompression into a lossy format. Every output is PNG so sharp boundaries and source transparency can be represented. A JPEG or WebP source is decoded and re-encoded, which means source metadata and original compression bytes are not retained even though the visible pixel region is preserved.
Inputs are limited to supported browser-decodable JPG, PNG, or WebP files up to 25 MiB. Decoded images must remain within 20,000 pixels per side and 40 megapixels total. Those limits protect the tab from unexpectedly large canvas allocations. Invalid types, empty files, decode failures, unavailable canvas support, invalid grid values, and failed PNG encoding produce a visible error rather than partial or falsely successful output.
Output names include the original base name plus row and column numbers, such as photo-row-2-column-3.png. This makes the files sortable and removes ambiguity when the last row or column is one pixel larger. The result also reports each tile's width, height, and approximate encoded size. Changing the source or either grid value invalidates old output, and temporary browser URLs are revoked when they are replaced or the tool closes.
This focused splitter processes one image at a time and creates a uniform row-by-column grid. It does not detect faces, automatically find panels, create overlapping slices, add bleed, package downloads as a ZIP, reconstruct a panorama, or optimize the PNG files. For print production, social-network templates, or grids with custom guides and gutters, use an editor that supports manual slice positions. For a straightforward nine-grid post, set three rows and three columns and download the nine numbered results.
Methodology & sources
Validate one bounded local image, calculate each grid boundary as floor(index times dimension divided by count), reject zero-pixel tiles, draw every source rectangle once onto its own natural-size canvas, encode PNG, preserve row-major naming, guard stale asynchronous work, and revoke replaced Object URLs.
Frequently asked questions
- Does the splitter lose pixels when dimensions are uneven?
- No. Integer boundaries cover every source pixel exactly once; uneven rows or columns differ in size by at most one pixel.
- Why are all tiles PNG files?
- PNG preserves exact tile boundaries and transparency without adding lossy JPEG compression to each extracted region.
- Is my source image uploaded?
- No. Decoding, slicing, encoding, and download creation run locally in the current browser tab.
- Can I split an image into a 3 by 3 Instagram-style grid?
- Yes. Set rows and columns to 3; the nine downloads are named by row and column in reading order.
Related tools
- Image CropperCrop images to any size or aspect ratio, right in your browser
- Circle CropCrop a JPG, PNG, or WebP into a centered circle locally and download a full-resolution transparent PNG.
- Image ResizerResize any image to exact pixel dimensions in your browser
- Combine ImagesArrange 2–12 local images horizontally, vertically, or in a centered grid and export one PNG.
- Add Background to PNGReplace PNG transparency with an exact solid background color locally while preserving the original pixel dimensions.
- Add Border to ImageAdd a solid 1–500 pixel border around a local JPG, PNG, or WebP and download a full-resolution PNG.
Image Tools guides
View all- How to Flip an Image Horizontally or Vertically
- How to Add Blur to an Image (Full or Partial)
- How to Crop Image in Canva: A Complete Guide
- How to Compress Image File Size in Your Browser
- What Colors Are Used in an Image: Pick Exact HEX Codes
- Create an Animated GIF in Photoshop or Your Browser
- How to Resize an Image Without Stretching or Losing Quality
- JPEG vs PNG vs WebP: Which Image Format Should You Use?
- How to Compress Images for the Web Without Losing Quality