Free Anti-AI Text Maker

People read the blur. Machines read the decoy. Type any message and this tool hides it inside a hybrid image, your real words become a soft ghost that human eyes read at a distance, while AI scrapers and OCR lock onto a sharp decoy made of lookalike letters. Download a flat png where your text never exists as data. No watermarks, no signup.

Squint or step back and the real message appears

Type a message, hide it in plain sight

Hit generate, then drag the zoom down and watch the decoy dissolve into your real words.

Zoom100%

Themes

Your own colors

BackgroundTextOutline

Zoom out until the outlines dissolve into the real words. Exports a flat png, so the message never exists as selectable text inside the file. It slows automated reading, it is not encryption.

People read the blur, machines read the decoy

This is a hybrid image, two messages living at two spatial frequencies. Your real words are rendered as a soft, bold ghost, the low frequency mass your visual system assembles the moment you squint, zoom out, or step back. On top sits a razor sharp decoy in concentric outline rings. Machine readers hunt edges, so they transcribe the decoy and walk away confident. Drag the zoom slider down and you can watch the handoff happen live.

Anti-AI text example zoomed out, the decoy dissolves and the hidden words read me irl appear in black on white

Every letter gets a lookalike decoy

The decoy is not random noise, it is engineered confusion. Every character in your message is swapped for a lookalike that shares its skeleton, A becomes 4, S becomes 5, M becomes N, so an edge focused reader lands on a confident wrong answer instead of a suspicious blank. Camouflage marks, stray dots, ticks, and arcs, add meaningless edges across the canvas to bury whatever signal is left. Make text anti-AI without changing a single word you wrote.

Text that AI cannot read, shown at a distance where the hidden message humans only becomes clearly readable

A flat png that never contains your text

The export is pixels, nothing else. Your message never exists as selectable text inside the file, so it cannot be copied, indexed, or scraped as data, it can only be seen. That makes it an anti-AI filter for your words, drop it into social posts, stream overlays, zine pages, or party invites where you want humans in on it and bots left out. Pick from sixteen themes, shuffle harmonized palettes, or dial in your own background, text, and outline colors, then print the squint-to-read effect big on custom stickers and posters. One honest note, the machine-fooling side is built for screens, a camera photographing a printed sign from a distance cannot capture the fine decoy detail, so in print this is a trick for human eyes.

Hidden message not for bots revealed by zooming out, the anti-AI decoy layer fades leaving readable black text

What is anti-AI text, and how the hybrid trick works

As scrapers, crawlers, and multimodal models read everything posted online, a countermovement is designing for human-only legibility. Artists run anti-AI filters over their work, and the same instinct powers the anti-AI fonts movement in typography. This tool applies it to your words, a message machines misread and people enjoy decoding.

Two messages, two audiences

Hybrid images exploit how vision works at different distances. Up close, your eye and any OCR engine feed on fine detail, the sharp decoy rings. Far away, detail collapses and only broad masses of light and dark remain, which is where your real message lives. The zoom slider simulates that distance, so you can tune exactly when the reveal happens.

Why machines fall for it

Text recognition pipelines are built to find high contrast edges and match them to letterforms. The decoy hands them exactly that, confident strokes shaped like the wrong characters, while the real message hides in soft gradients they are trained to ignore as background. The scattered camouflage marks push the noise floor higher still.

Honest about the limits

This is obfuscation, not encryption. Anyone holding the image can zoom out and read it, that is the fun. What it resists is bulk automated reading, the silent indexing of usernames, addresses, and plans inside screenshots. It is also a screen technique, print a sign and any camera far enough away loses the decoy detail, leaving only the human-readable blur. For the physical version, put it on die cut stickers or business cards people have to step back from.

Frequently asked questions

Yes. Typing, styling, and downloading your image is completely free. Everything runs in your browser, so your message is never uploaded, and there is no watermark, signup, or limit on how many you can make.

Anti-AI text is a message designed to be read by people but resisted by machines. This tool builds it as a hybrid image, your real words become a soft blurred layer that human eyes assemble at a distance, while a sharp decoy message sits on top for edge-hungry scrapers and OCR to latch onto.

Three layers work together. Your real message is rendered as a soft ghost that carries the low frequency shape of the words. A decoy message, built by swapping every letter for a lookalike character, is drawn as sharp concentric outlines. Camouflage marks add meaningless edges on top. Machine readers follow the sharp edges and transcribe the decoy, people squint and read the ghost.

Zoom out, squint, or step back from the screen. As the fine detail in the outlines collapses, the blurred mass of the real words takes over and the message appears. The zoom slider under the preview does exactly this, drag it down until the decoy dissolves.

It slows automated reading, it is not encryption. The exported file is a flat png, so your message never exists as selectable text inside it, and standard OCR locks onto the decoy characters. A determined person with the image can still zoom out and read it, which is the point, it filters machines, not humans.

For people, yes, and it is honestly even more fun in print. For fooling AI, no. The decoy only works when a machine reads the image at full resolution, pixel for pixel, which is what happens on screens. Photograph a printed sign from across the room and the camera cannot resolve the fine decoy detail, all that survives is the soft blur of your real words, exactly what a person sees, so a capable AI reading that photo reads what a human reads. Treat print as the squint-to-read party trick and screens as the protection.

No. This tool never rewrites your words. AI humanizers change what your text says so it passes detectors. This does the opposite job, it protects how your text is read, wrapping your exact message in a layer that machine readers misread.