Add Text to Image

Several pieces of text, each placed where it belongs

This is the part most free tools get wrong. They give you one text box, or a fixed caption slot at the top and bottom, and anything more complicated than that is out of reach.

Here you add a layer, drag it where it should sit, and add another. Which covers the things people actually need: a caption along the bottom and a label pointing at something in the middle; a title plus a small credit in the corner; three labels naming three parts of a diagram; a quote with an attribution underneath it.

Click a layer to select it, and the options then edit that layer — so each one can have its own size, colour and font. Rotating the image re-anchors every layer through the rotation, so a label you placed on someone's head stays on their head rather than sliding off into the background.

The outline is what makes it readable

White text with a black outline is not only a meme convention; it is the practical solution to a real problem. Photographs are busy and uneven, and plain white text vanishes wherever the picture happens to be pale — a sky, a shirt, a window. The outline guarantees the words stay legible wherever they land.

So leave the outline on for anything over a photograph. There is a No outline option for when the background is genuinely plain — a flat colour, a clean product shot on white, a screenshot of a document — where an outline just looks clumsy. Shadow is the softer alternative: less assertive than an outline, and often better for a title that is meant to look designed rather than stamped on.

Opacity down at 30–50% turns a text layer into a watermark rather than a caption, which is useful when you want your name on an image without it competing with the picture.

Choosing a font that suits the job

Eight fonts, and the default is Impact — which is the meme typeface. That is right if you are making something jokey and wrong if you are not, so it is the first thing to change for a plain caption or a professional label.

Arial or Verdana for a neutral caption or a label on a screenshot; Verdana is the more legible of the two at small sizes. Georgia or Times New Roman for a quote or anything that should read as considered rather than casual. Courier New for something technical. Comic Sans when you mean it.

These are the fonts every device already has, which is why they work without waiting for anything to load — but it also means genuinely distinctive typography is not on offer here. For brand typefaces you would need a design tool.

Which format to save as

The answer follows from what is underneath the text, and it is the same rule as everywhere else on this site.

PNG is the default and correct whenever the image contains sharp edges — screenshots, diagrams, anything with existing text or a flat background. JPEG compression puts grey fuzz around every hard edge, which is exactly what your new text is made of.

JPG is right for a photograph headed somewhere size matters. The text on it will pick up a little compression at the edges, which on a photographic background is invisible at normal quality.

WebP is the modern choice for the web: much smaller than PNG at similar quality, and supported everywhere current. Not the choice if the file has to go into older desktop software or a print workflow.

Nothing is uploaded

Which matters more for this tool than it sounds, because of what people put text on. Screenshots of conversations with names in them. A photo of a document being annotated for a colleague. A picture of a child for a birthday card. None of that needs to travel to a server in order to have three words placed on it.

Related tools

Meme Generator does the fixed top-and-bottom meme layout if that is all you need. Add Watermark to Image repeats a mark across an image or tiles it, which is the right tool for protecting photos rather than captioning them, and Crop Image tightens the framing first — usually a bigger improvement than any text setting.