About ASCIILogo.com

Built for logos, not photographs

Most converters treat every image the same way: sample a grid of pixels, map brightness to a ramp of characters, done. That is fine for a portrait, where the eye forgives a lot. A logo is mostly hard edges and flat colour - precisely what naive sampling destroys first. Three things here are done differently, and together they are the whole difference:

Properly filtered, not point sampled

The image is resampled with a real Lanczos filter down to exactly one pixel per character cell, so every character represents a correctly averaged area of your logo. This is what makes the tone smooth instead of noisy.

Round things stay round

Monospace characters are taller than they are wide. The cell aspect is measured from the font actually rendering on your screen rather than assumed, so circles come out circular and your mark keeps its proportions.

Edges stay sharp

A Sobel pass finds the real outlines and draws them with directional characters, so the silhouette that makes a logo recognisable survives instead of dissolving into the grey ramp.

Start from anything

Four ways in

Drag and drop, click to browse, paste straight from the clipboard, or give it an image address. You can also drag an image directly off another web page.

Every common format

PNG, SVG, JPG, GIF and WebP. Transparent backgrounds are understood as transparency, not as white, so a knockout logo converts the way you expect.

Vector stays sharp

SVGs are rasterised at high resolution before sampling. A vector logo with a tiny declared size converts from real detail instead of from a postage stamp.

A whole folder at once

Batch mode converts multiple files in one pass and downloads the lot - useful when you have a brand family rather than a single mark.

Fix the framing first

Crop, rotate and flip before conversion, and trim empty margins automatically so the logo fills the frame instead of floating in whitespace.

Tune it until it looks right

Grids up to 1000 x 1000 characters

A million cells if you want them. The character grid is the primary control and character size turns it into pixels, with the exported dimensions shown live as you move either one.

Any character set, including your own

Seven built-in ramps from a coarse three-step to a 70-level detailed ramp, plus blocks, dots and binary. Or type your own string - your initials, your brand letters, anything, multi-byte characters included.

Dithering for smooth gradients

Floyd-Steinberg error diffusion stops gradients banding into ugly steps at low detail. It is the difference between a logo that looks deliberate and one that looks like a fax.

Outline mode

Draw only the edges and leave everything else blank, for a clean line-art take on the mark.

Real colour, or none

Keep your logo's own colours per character, quantised to true colour, 256 colours or the ANSI 16 palette for terminals. Or go monochrome and pick your own ink and background, including fully transparent.

Typography matters

Choose the font family, line height and letter spacing. Font choice changes the result more than most of the sliders do, and the geometry re-measures itself when you change it.

The usual tone controls

Brightness, contrast, gamma and edge sensitivity, each with a numeric readout so you can get back to where you were.

Export to where it is actually going

ASCII art almost always has a destination. Rather than handing you one format and leaving you to rebuild the rest by hand, it writes all six:

PNG

The rendered image, at the exact size you set, with an optional transparent background.

Plain text

Just the characters, trimmed tight to the artwork.

Markdown

A fenced code block ready to paste into a README, with the fence lengthened automatically if your art contains backticks.

HTML

A self-contained coloured block you can drop straight into a page - colour runs are grouped, so it stays small instead of emitting a tag per character.

ANSI

A terminal banner with real escape codes, in true colour, 256 colours or the ANSI 16 palette, depending on what your terminal supports.

SVG

Vector output that scales to any size without resampling - print it, put it on a billboard, it stays crisp.

Made to be used more than once

Presets for real destinations

An 80-column terminal, a README block, a square avatar, a 1920 x 1080 wallpaper. Each sets size, detail, character set and colour together, and everything stays editable afterwards.

It remembers, and it shares

Your settings persist between visits and encode into the address bar, so you can send someone a link to the exact configuration that produced a result.

See the difference

A before and after view puts the original next to the conversion, so you can judge whether it still reads as your logo - which is the only question that matters.

Big grids stay usable

A 1:1 zoom with a scrollable stage lets you inspect a thousand-character grid at actual size instead of squinting at a thumbnail.

Recent conversions and shortcuts

Jump back to something you made earlier, and drive the whole thing from the keyboard.

Fast, private and accessible

It keeps up with you

A full 1000 x 1000 render - a million characters, with edge detection, dithering and colour all enabled - completes in about 186 milliseconds. At everyday settings it is effectively instant, so dragging a slider shows you the result as you move it.

Private by construction, not by promise

The conversion happens entirely in your browser. Nothing is uploaded, and the site's content security policy forbids outbound connections altogether - so "your image never leaves this page" is enforced by the browser rather than by a privacy policy you have to take on faith.

Readable by screen readers

The finished art is exposed as real text, not only as pixels on a canvas, so it can be read aloud, selected and copied.

Works on a phone

The full control set and the zoomable preview both work on a small screen.

Why I built this

Most ASCII art converters treat every image like a photograph. They sample a grid of pixels, map brightness to a ramp of characters, and hand you the result. That works for a portrait, where the eye forgives a lot. It falls apart on a logo, because a logo is mostly edges and flat colour - the two things naive sampling destroys first. You end up with a grey smear that technically encodes the image and looks nothing like the mark.

I wanted one that respected the logo. Three decisions do most of that work. The image is resampled with a proper Lanczos filter down to exactly one pixel per character cell, so each character is a correctly filtered average instead of a point sample. The character cell aspect is measured from the font actually in use rather than assumed, which is why circles come out round instead of squashed. And a Sobel pass finds real edges and draws them with directional characters, so hard outlines stay hard instead of dissolving into the ramp.

The other half was practical. I did not want to upload a client's logo to somebody else's server just to get a text version of it, so the whole thing runs locally in the browser and the site is not permitted to make outbound connections. And ASCII art is almost always going somewhere specific - a README, a terminal banner, a web page - so it exports in those formats rather than making you rebuild the output by hand.

It is free, there are no accounts, and there is nothing to install.

About RJ Lindelof

Technology executive - 35+ years in software development - 21+ years in engineering leadership

I have spent decades building enterprise-scale systems and leading global development teams, and somewhere in all of that I kept running into the same thing: the small utility that wants a server round trip, an account, a cookie banner or your data before it will do a ten-second job. I have been the engineer who lost an afternoon to that, and the executive who had to justify the spend to a board. I speak both languages fluently.

So the rule here is simple - the tool should respect the person using it. ASCIILogo.com runs entirely in your browser. It stores nothing, sends nothing, and once the page has loaded, converting an image needs no network at all. No account, no paywall, no consent dialog standing between you and a logo you wanted rendered thirty seconds ago. Speed, privacy and getting out of your way drive every decision in it.

This is the tool I wanted years ago and could not find: practical, honest, and free of the usual friction. It belongs to a wider set built on the same idea - AI-focused development resources at RJL.ai, the broader technology community at RJL.dev, and everything else I build collected at RJL.link.

If it is useful to you, or it mangled a logo it should have handled, I would like to hear about it.