Photo as text. Text as art.

No. The conversion runs in your browser — pixel sampling, character mapping, canvas rendering, all client-side. Your photo never touches our servers — we don't run an image pipeline at all.

Light (10 chars) is the classic ASCII look — clean, readable, fits 80-column terminals. Dense (70 chars) gives you near-photographic detail at high resolutions. Block (unicode shading) is best for portraits because the characters fill space more evenly. Binary (0/1) is the bulletin-board nostalgia option.

The text version is just N×M characters — small. The PNG renders each character at a chosen pixel size, so a 200-column ASCII at 12px characters becomes a 1500-pixel-wide image. Drop columns down or the rendered character size if you need a smaller PNG.

We resize the photo down to a tiny grid (one cell per character), measure the average brightness of each cell, and pick a character based on that brightness. Dark cells get dense characters (like @ or #); light cells get sparse ones (like . or space). Reverse with the Invert toggle for light backgrounds.

Yes — toggle Custom and type any string. Order matters: dark to light, left to right. Try ' .o0@' for a quick custom 5-char set, or ' ░▒▓█' for unicode block shading. Empty space at the start makes light areas blank.

ASCII art shows brightness, not detail. Faces with strong tonal contrast (clear shadows on the cheek, lips, hair) read well. Flat-lit selfies often come out muddy. Tip: bump the original through Photo Filters (Punchy preset) before converting.