Definition

What is pixel art?

Pixel art is a picture drawn on a coarse grid, where every square is one solid color and you can count the squares.

A 32×32 icon has 1,024 squares. A phone photo might have two million. The craft is deciding which squares survive. Console hardware used to make that decision for you. An NES picture is 256×240, and a lot of sprites were stored as 8×8 or 16×16 tiles because that's what the cartridge held. People kept the look after the limits went away, because a counted grid is easier to edit, to animate, and to copy onto beads or stitches.

What a converter actually does

An image to pixel art converter doesn't invent a drawing. It resamples your file onto a width you pick, then snaps the colors to a shorter list. Set the width to 32 and you get 32 columns. You can count them. If the source is 1200×900 and the width is 40, the height lands near 30, and each cell covers about 30×30 source pixels. Eyelashes disappear. The hair mass stays. That trade is the whole tool.

Three numbers matter more than any preset name.

• Width, in cells. 16 reads as an icon. 32 can hold a character. 64 starts to look like a simplified photo.
• Color count. 4 colors is a handheld-console look. 16 matches a Pico-8 game. 32 is plenty for a portrait once the grid is already coarse.
• Dithering. Off means flat areas of one hex. Floyd–Steinberg mixes neighbors so a sky doesn't turn into three stripes. Atkinson keeps more contrast and leaves more untouched squares.

What it's not

It's not a text-to-image model. You bring the picture. The page only changes the grid and the palette. It's also not a guarantee that the result is "good art." A busy street photo at 16 pixels wide becomes mud. A logo with a clear silhouette often looks fine at 32. If the result is noisy, raise the pixel width or cut colors. The muddy conversion guide walks through that order.

A social-app pixelate slider is a different operation. It averages blocks and then scales them back up with soft edges. You can't point at a square and say it's #e4572e. A real grid can. The longer comparison is in pixel art vs a pixelate filter.

What you can do with the grid

The same cells become different files. A small PNG at 32×32 is a sprite. A large PNG, scaled with hard edges, is something you can look at without squinting. A symbol chart is a cross-stitch or fuse-bead plan. Row and column counts from a 2-color grid are Nonogram clues. Pixelate Tools is that one pipeline with several exits. Start at the image to pixel art converter, then move to the sprite converter if you need transparency and a game-sized grid.

You still need a source you're allowed to use. Processing happens in the browser. The copyright on the photo doesn't fall off because a tool resampled it. If the picture is yours, the PNG is yours too.

If you only remember one check: zoom in until you see squares, and confirm each square is a single color. If the edges are fuzzy, you exported a preview, not the pixel grid.

Mistakes that look like pixel art

Exporting the large preview and calling it a sprite is the usual one. Re-uploading that file runs the tool on squares, not your photo. Dithering on for a craft chart looks rich on screen and turns the bead board into a checker you'd place by hand.

Checklist before you share a file

Open the small PNG. Dimensions should match the meta line. Legend length should match the Colors slider unless a fixed palette is selected. Zoom to 800%: edges should stair-step, not blur. If the goal is a game sprite, confirm transparency in the small file on a colored background. If the goal is stitch or beads, confirm dithering is off and read the sibling walkthrough for that exit. The photo walkthrough and pixel converter are the fastest way to produce a grid you can defend in one sitting.

Questions

Keep going

• Turn a photo into pixel art
• Pixel art vs a pixelate filter
• 16, 32, and 64 pixel sprites
• Open the converter

Updated 30 September 2026