Which photos work best?

Photos where the subject has a clean outline and some contrast with the background — portraits, pets, products, buildings against the sky. Place the text behind the upper part of the subject so the head or shoulders cover a small part of it: the depth effect is obvious and the words are still readable. Tiny subjects or very busy backgrounds can give rougher cutout edges.

Why does the first run take a while?

The cutout uses the ISNet segmentation model, about 46 MB. It’s downloaded once from this site and cached in your browser, so later uses start immediately. Segmentation runs on your own device and is faster in browsers with WebGPU — usually a few seconds.

Can I use my own fonts?

The font list uses fonts that ship with your operating system, so the exact look can differ slightly between computers: Impact, Georgia and Arial-style fonts exist on both macOS and Windows, while Chinese fonts like Kaiti depend on the system. Missing fonts fall back to a similar one automatically, and the export always matches the preview.

Is my photo uploaded?

No. Both the AI model and the text compositing run locally in your browser, so the photo never leaves your device. Exports are redrawn at the original resolution with no quality loss; when saving as JPG, any transparent areas are filled with white.

🧍

🧍Text Behind Image - Put Text Behind a Person or Subject

Create the magazine-cover look where a headline sits behind the person in the photo — no Photoshop needed. When you upload an image, the ISNet segmentation model runs in your browser and separates the subject (a person, pet, product or building) from its background, so the picture is rebuilt as three layers: background, your text, then the subject on top. Add as many text layers as you like and style each one with a system font (Impact, sans-serif, serif, rounded, script, Kaiti and more), weight, size, colour, stroke, shadow, opacity, rotation and letter spacing, then drag it into place right on the preview. Any layer can also be moved in front of the subject for a layered, overlapping look. Give the subject an outline or a soft glow so it pops out from the lettering. Exports are redrawn at the photo’s original resolution as PNG or JPG, and both the photo and the cutout stay on your device.

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

Loading tool...

How to use Text Behind Image

1Upload a photo with a clear subject and wait for the AI to find it (the 46 MB model downloads once on first use)
2Edit the text, font, colour and stroke, drag it into place on the preview and add more layers if you like
3Optionally add a subject outline or glow, pick PNG or JPG and export at full resolution

FAQ

Which photos work best?: Photos where the subject has a clean outline and some contrast with the background — portraits, pets, products, buildings against the sky. Place the text behind the upper part of the subject so the head or shoulders cover a small part of it: the depth effect is obvious and the words are still readable. Tiny subjects or very busy backgrounds can give rougher cutout edges.
Why does the first run take a while?: The cutout uses the ISNet segmentation model, about 46 MB. It’s downloaded once from this site and cached in your browser, so later uses start immediately. Segmentation runs on your own device and is faster in browsers with WebGPU — usually a few seconds.
Can I use my own fonts?: The font list uses fonts that ship with your operating system, so the exact look can differ slightly between computers: Impact, Georgia and Arial-style fonts exist on both macOS and Windows, while Chinese fonts like Kaiti depend on the system. Missing fonts fall back to a similar one automatically, and the export always matches the preview.
Is my photo uploaded?: No. Both the AI model and the text compositing run locally in your browser, so the photo never leaves your device. Exports are redrawn at the original resolution with no quality loss; when saving as JPG, any transparent areas are filled with white.