Which method should I use?

For detailed textures without a strong direction — stone, grass, sand, fabric — patch-based healing usually looks the most natural: it covers the seam with similar areas from the image and keeps the size. Cross-fade is fine for fine-grained, gently varying textures, but large repeated objects may show see-through ghosts. Mirror tiling is guaranteed seamless if you don’t mind symmetry; the tile becomes twice as wide and tall.

Why check the offset-by-half view?

When a tile repeats, the joins are its left/right and top/bottom edges. Shifting the image by half its width and height moves those edges to the centre as a cross, where problems are easiest to spot. If no cross is visible after processing, the tile joins seamlessly in both directions.

What does the seam visibility score mean?

It divides the average colour difference across the wrap-around edges (last column vs first column, last row vs first row) by the average difference between neighbouring pixels inside the image. A value near 1 means the joins change no more than any two adjacent pixels, so they’re invisible; an ordinary photo tiled as-is typically scores between 5 and 20.

Will resizing on export bring the seams back?

No. The tool lays the tile out 3×3 before scaling and keeps the middle copy, so edge pixels are interpolated with their wrap-around neighbours and the tile still joins seamlessly at any size.

♾️

♾️Seamless Texture Maker - Make Any Image Tile Seamlessly

A seamless (tileable) image repeats left-to-right and top-to-bottom without visible joins, which is what you need for website backgrounds, 3D materials, fabric and packaging prints. This tool offers three ways to turn a photo or texture into one. Mirror tiling is the simplest: the image plus three mirrored copies form a 2×2 tile that is always seamless, at the cost of obvious symmetry. Cross-fade blending fades a strip from the right and bottom edges over the left and top, which suits fine, even textures. Patch-based seam healing borrows from the image quilting algorithm: the image is offset by half so the seams meet in the middle, then the seam is covered with best-matching patches taken from the image itself, each joined to its neighbours and to the original along a minimum-error boundary cut, so the size stays the same and there’s no ghosting. An “even out lighting” option removes large brightness or colour gradients first. Previews include offset-by-half with the old seams marked, a before view and 2×2, 3×3 or 4×4 tiling with zoom, plus a seam visibility score. The second mode repeats a motif (ideally a transparent PNG) in a grid, half-drop or brick layout with cell size, motif size, rotation and alternate flipping, producing a ready-to-repeat tile. Export at any pixel size — resampling wraps around, so the result stays seamless — as PNG or JPG. Everything runs in your browser and nothing is uploaded.

Free · No sign-up · Runs in your browserLast updated: 2026-10-26

Loading tool...

How to use Seamless Texture Maker

1Choose “Make a photo / texture seamless” and drop an image, or “Repeat a motif” and drop a motif
2Pick a method (seam healing, cross-fade or mirror), tune it and check the seams in the offset and tiled previews
3Set the export size and format, then download the tile

FAQ

Which method should I use?: For detailed textures without a strong direction — stone, grass, sand, fabric — patch-based healing usually looks the most natural: it covers the seam with similar areas from the image and keeps the size. Cross-fade is fine for fine-grained, gently varying textures, but large repeated objects may show see-through ghosts. Mirror tiling is guaranteed seamless if you don’t mind symmetry; the tile becomes twice as wide and tall.
Why check the offset-by-half view?: When a tile repeats, the joins are its left/right and top/bottom edges. Shifting the image by half its width and height moves those edges to the centre as a cross, where problems are easiest to spot. If no cross is visible after processing, the tile joins seamlessly in both directions.
What does the seam visibility score mean?: It divides the average colour difference across the wrap-around edges (last column vs first column, last row vs first row) by the average difference between neighbouring pixels inside the image. A value near 1 means the joins change no more than any two adjacent pixels, so they’re invisible; an ordinary photo tiled as-is typically scores between 5 and 20.
Will resizing on export bring the seams back?: No. The tool lays the tile out 3×3 before scaling and keeps the middle copy, so edge pixels are interpolated with their wrap-around neighbours and the tile still joins seamlessly at any size.