Edit text in a JPG when the layers are gone.

That is why clicking a word in a JPG does not reveal a font, text box, or cursor. The file no longer knows which pixels form the text and which form the background. It only stores the finished image. If you still have the original Canva, Photoshop, Figma, PowerPoint, or other editable source, change the live text there and export a fresh JPG. That gives you the most exact result and keeps the design editable for next time.

If the source file is gone, replacing the wording becomes an image repair. The old letters need to be removed, the hidden background needs to be rebuilt, and the replacement needs to match the nearby perspective, color, lighting, and texture. A focused image editor can generate that small region while leaving the rest of the JPG alone.

ImageTextEdit is for the first job: replacing text that is already baked into a finished image. It is not an OCR tool, and it is more than a text box placed over the old wording. The selected area is rebuilt so the replacement can sit in the scene instead of floating above it.

This works best for one contained change at a time. A wrong menu price, an expired event date, a misspelled sign, or a short product label is a better fit than a page of fine print. For a broader explanation of how the pixels are rebuilt, read how AI image text editing works.

1. 1

   Upload the best JPG you have

   Use the original export rather than an image copied from a chat thread or social feed. Extra compression can soften letter edges and reduce useful texture.

2. 2

   Select the complete text region

   Draw around every letter plus its outline, glow, shadow, or reflection. Include a narrow edge of the immediate background, but keep unrelated objects outside.

3. 3

   Enter only the replacement text

   Type the wording exactly as it should appear. Check capitalization, punctuation, currency symbols, and spacing before generating.

4. 4

   Generate and compare with the source

   Inspect the result at full size. If the spelling, baseline, texture, or nearby detail is wrong, adjust the selection or wording before trying again.

You can prepare the whole edit in the image text editor before generation starts. The blank upload remains the main path, and no credit is used merely for uploading, selecting, or typing the replacement.

JPEG compression can create small blocks or halos around high-contrast letters. Those artifacts are part of the source pixels, so inspect the repair at the same zoom where the finished asset will be used. A result that looks natural in a web card may not be suitable for a large print.

• Read every letter, number, and symbol at full size.
• Compare the baseline, spacing, angle, and scale with nearby text.
• Look for smears, repeated texture, or remnants of the old wording.
• Check that borders, objects, and labels outside the selection stayed intact.
• Keep the original JPG so the change can be reviewed or reversed.
• Confirm you own the image or have permission to update it.

For a step-by-step workflow that also covers PNG and WebP files, see how to edit text in a flattened image without Photoshop.