Text Behind Image

How to Put Text Behind an Image

The page opens with a demo photo already loaded — our friend the farmer, with "it ain't much" tucked neatly behind his cap — so you can see exactly what the effect looks like before you touch anything. When you're ready, swap in your own picture:

1. Click "Upload your photo" and pick an image with a clear subject. A person works best, but a dog, a car, a coffee cup — anything that stands apart from its background will do.
2. Type your text. Hit Enter if you want a second line. The text slides in behind your subject automatically; there's no masking, no layers panel, nothing to cut out by hand.
3. Drag it into position right on the preview. Grab the canvas and pull the text wherever it reads best — usually so the subject overlaps a letter or two. That overlap is what sells the depth.
4. Style it. Pick a font, adjust the size and letter spacing, add an outline, rotate it, drop the opacity, or tilt it in fake-3D if you're feeling fancy.
5. Download your PNG. The export comes out at your photo's original resolution, not the preview size — so a 4000px photo gives you a 4000px result.

What's Actually Happening Under the Hood

When your photo loads, a small AI segmentation model runs directly in your browser and figures out which pixels belong to the subject and which belong to the background. The tool then rebuilds your image in three layers: the original photo at the bottom, your text in the middle, and a cutout of the subject pasted back on top. Your eye reads the stack as depth — the text looks like it was standing in the scene all along.

Because all of this happens on your device, there's no upload queue and no waiting for a server to process anything. The first photo takes a few extra seconds while the model downloads; after that, it's nearly instant. Your picture never leaves your browser during editing, which also means it works fine on photos you'd rather not send to some random server.

Where You've Seen This Effect Before

Everywhere, honestly. YouTube thumbnails lean on it constantly — a creator's head popping out in front of a giant "EXPOSED" or "I TRIED IT" is basically the house style of the platform at this point. Spotify-style lyric posters use it. Album art uses it. Instagram photographers use it for birthday posts, travel recaps and engagement announcements. Sports pages put the athlete's number behind them. Streamers put their handle behind their avatar.

The reason it works is simple: overlapping is one of the strongest depth cues your brain has. When a person's shoulder covers part of a letter, the text stops being a flat caption slapped on top and becomes part of the scene. Flat text competes with your subject for attention; text behind the subject frames them instead. Same photo, same words, completely different feel.

Choosing a Photo That Cooperates

The segmentation model is good, but you can make its job easy or hard. Photos where the subject is clearly separated from the background — different colors, decent lighting, a bit of distance — produce crisp cutouts with clean edges. Photos where a dark jacket melts into a dark doorway will give you softer, mushier edges around the overlap.

A few things that consistently help:

• Contrast between subject and background. A person in a light shirt against trees beats a person in green against the same trees.
• One obvious subject. Group shots work, but the effect reads strongest when a single person or object owns the frame.
• Some headroom or side room. You need somewhere for the text to live. Photos cropped tight to the face leave nowhere to put it.
• Sharp focus on the subject. If the photo already has natural background blur, even better — the text inherits that layered, photographic feel.

Portrait shots taken on a phone in decent light are the sweet spot. That's most of what people feed this tool, and it's what the model was trained to be best at.

Getting the Typography Right

Half-hidden text has to survive losing 30–40% of itself, so heavy fonts win. Anton is the default here for a reason: it's thick, tall and reads clearly even when a whole shoulder is parked in front of it. Bebas Neue and Archivo Black are close cousins. Impact is the classic meme choice. If you want something softer, Pacifico and Caveat give you a handwritten look — just make them bigger to compensate for the thinner strokes.

There are twelve curated display faces up front and a searchable catalog of about a hundred more Google Fonts behind the search box, loaded on demand. Practical advice: pick your font before you fine-tune the position, because switching from a condensed face to a wide one moves everything.

For color, white is the safe default against most photos. But sampling a color that's already in the scene — the sky, a shirt, a neon sign — is the trick that makes a result look designed rather than generated. Use the color picker, aim for something in the background, then nudge the opacity down to 85–90% so the text settles into the light of the photo.

The Small Adjustments That Sell It

Outline saves you when the text crosses both bright and dark regions of the photo. A few pixels of dark stroke keeps letters legible where they pass over a bright sky. Set it to zero when your background is even.

Rotation and 3D tilt are easy to overdo. A couple of degrees of rotation adds energy; forty-five degrees adds chaos. The tilt sliders shear the text plane so it appears to recede into the scene — subtle values (10–20) mimic the perspective of the photo surprisingly well.

The flip option inverts the whole trick: instead of the subject sitting in front of the text, the background does, and the text appears pinned to your subject. It's the right choice maybe one time in ten, but when it's right, it's very right — try it when the auto-cutout grabs the "wrong" half of your scene.

Add a Watermark Before You Post

There's a dedicated watermark field below the styling controls. Type your handle or domain, then place it with the position presets (corners, edges, center) or drag the X/Y sliders for an exact spot. You can set its size in pixels, pick a color, and toggle a contrast stroke so it stays readable on any background. The watermark is baked into the downloaded PNG itself — so when your image gets reposted, and the good ones always get reposted, your name travels with it.

Share a Link, Not Just a File

The Share button creates a link that carries your whole composition — photo, text, styling and watermark settings. Anyone who opens it lands on this page with your exact setup loaded, ready to tweak or download. It's handy for getting a second opinion before posting, or for handing a template to a teammate: set up the style once, share the link, let them swap the photo.

Frequently Asked Questions

Do I need Photoshop for this?

No. What you'd do in Photoshop with subject selection, layer masks and a text layer, this does in one step in the browser. Photoshop still gives you more control for surgical edits, but for the standard text-behind-subject look, there's no meaningful difference in the result — just about ten minutes of difference in the effort.

Is it free?

You can build your composition and preview everything without an account. Downloading the full-resolution PNG requires signing in, and depending on the current promotion it's either free for a limited time or included with a subscription — the badge at the top of the tool shows you which applies right now.

What file types can I upload?

JPG, PNG, WebP, GIF, AVIF, BMP and HEIC/HEIF straight off an iPhone. The export is always PNG at the original resolution of your photo.

Why did the cutout miss part of my subject?

Usually low contrast — the subject's edge blends into the background, or the lighting is very dim. Try a photo where the subject stands out more, or use the flip option, which sometimes produces a cleaner split on tricky images. Busy scenes with many overlapping people are the hardest case.

Can I write multiple lines?

Yes — press Enter in the text box. Lines are centered as a block, and the size slider scales them together. Two short lines usually beat one long one, since each line gets its own chance to overlap the subject.

Does it work on my phone?

Yes. The segmentation runs fine on modern phones, and dragging the text around with your thumb is honestly more natural than with a mouse. Downloads land in your camera roll or files app like any other image.