How to Remove Text from an Image: 6 Methods Ranked

You have a photo with text stamped on it — a watermark on an image you own, a caption baked into a screenshot, a date stamp in the corner of a vacation photo — and you need a clean version. The text has to go, the background behind it needs to look natural, and you do not want to spend an hour doing it.

The right method depends on the complexity of what is behind the text. Solid-color backgrounds are trivial. Complex textures (grass, fabric, crowds) require smarter inpainting. This guide covers six real methods — desktop software, browser tools, and mobile apps — with honest accuracy assessments so you pick the right one the first time.

Important note on watermarks: Removing watermarks from images you do not own or license is copyright infringement. The techniques below are for removing text from your own images — watermarks on photos you shot, overlays you added, captions from your own screenshots, or date stamps from your own camera.

Method Comparison

If you are working with screenshots specifically and want to crop the text out rather than inpaint it, cropping the image is faster and produces a cleaner result when the text sits at an edge.

Method 1: Photoshop Content-Aware Fill (v26.3)

Photoshop's Content-Aware Fill remains the most accurate tool for removing text from complex backgrounds. Version 26.3 (the current Creative Cloud release) uses Adobe's Firefly-powered generative fill, which reconstructs textures — brick patterns, foliage, fabric weave — with near-perfect continuity.

Steps

1. Open the image in Photoshop v26.3.
2. Select the text. The Object Selection Tool (W) works if the text has clear contrast. For scattered text, use the Lasso Tool (L) and draw around each text block. Hold Shift to add multiple selections.
3. Expand the selection by 2-4 pixels: Select > Modify > Expand > 3 px. This captures the anti-aliased edges that a tight selection misses.
4. Open Content-Aware Fill: Edit > Content-Aware Fill. The preview panel shows the result before you commit.
5. Adjust the sampling area (green overlay) to exclude other text or elements you want to preserve. Shrink it if the fill is pulling from the wrong area.
6. Click OK. The text disappears and the background reconstructs.

When to use Photoshop

• The background behind the text is complex — patterned fabric, a crowd, textured wall
• You need pixel-perfect results for print or commercial use
• You are removing text from multiple images (Actions can automate the selection + fill workflow)

When to skip Photoshop

• You do not already have a Creative Cloud subscription ($22.99/mo is steep for occasional use)
• The background is simple enough that a free tool handles it fine

For the reverse operation — pulling text out of an image instead of erasing it — see our guide on extracting text from images with OCR.

Method 2: GIMP Heal Tool (v2.10.38)

GIMP is the free alternative that handles most text removal tasks. The Heal tool (introduced properly in the 2.10 series) samples a nearby texture and blends it over the text, accounting for lighting and color gradients. It is not as sophisticated as Photoshop's generative fill, but for most backgrounds it produces clean results.

Steps

1. Open the image in GIMP v2.10.38.
2. Select the Heal Tool from the toolbox (or press H).
3. Hold Ctrl and click on a clean area near the text — this sets the source texture.
4. Paint over the text. GIMP blends the source texture into the target area, matching brightness and color.
5. For large text blocks, work in small strokes rather than one long drag. Reset the source point (Ctrl+click) frequently to avoid repeating patterns.

Alternative: Clone Stamp for solid backgrounds

If the background behind the text is a solid or near-solid color, the Clone Stamp (C) is faster than Heal. Set the source on the clean background, paint over the text, done. The Clone Stamp copies pixels exactly — no blending — which is perfect when there is nothing complex to blend.

Tips for clean results in GIMP

• Zoom in. Work at 200-400% zoom. Text removal at 100% zoom leaves visible artifacts you will only notice later.
• Use a soft brush. Hard edges on the heal brush create visible boundaries. Set hardness to 50-70%.
• Work in layers. Duplicate the background layer first (Layer > Duplicate Layer). Heal on the copy. If the result looks wrong, delete the layer and start over instead of stacking undo operations.

GIMP is licensed under GPL-3.0. The v2.10.38 release is the latest stable as of early 2026. GIMP 3.0 (in release candidate as of this writing) brings non-destructive editing, but the Heal tool workflow is identical.

Method 3: Canva Magic Eraser

Canva's Magic Eraser is built into the image editor — brush over the text, and Canva's AI reconstructs the background. It works directly in the browser and on mobile, which makes it the fastest option if you already use Canva for design work.

Steps

1. Upload the image to Canva (canva.com) or open it in the Canva mobile app.
2. Click Edit Image in the toolbar.
3. Select Magic Eraser from the effects panel.
4. Brush over the text you want to remove. Adjust brush size with the slider.
5. Canva processes the removal and shows the result. Brush again if remnants remain.

Limitations

• Free tier: Magic Eraser is available on the free plan with limited monthly uses. Heavy usage requires Canva Pro ($12.99/mo).
• Complex backgrounds: Canva's inpainting struggles with highly textured or patterned backgrounds. It performs well on gradients, skies, and solid colors but can produce smudged results on detailed textures like brickwork or foliage.
• Resolution: Canva compresses images. If you need the original resolution preserved, export at the highest quality setting and compare.

Magic Eraser works well for social media graphics where the text is on a designed background you control. For photographs with complex scenes, Photoshop or Cleanup.pictures produces better results.

Method 4: Cleanup.pictures (Browser)

Cleanup.pictures is a browser-based inpainting tool that punches above its weight. Upload an image, brush over the text, and the AI fills in the background. No account required for basic use.

Steps

1. Go to cleanup.pictures in your browser.
2. Upload or drag your image onto the canvas.
3. Adjust the brush size to match the text height.
4. Paint over the text. The tool processes each stroke and shows the result immediately.
5. Download the cleaned image.

Why it works well

Cleanup.pictures uses a neural inpainting model that handles both simple and moderately complex backgrounds. Grass, sky, walls, fabric — it reconstructs these textures convincingly. It processes the image server-side, so your local hardware does not matter.

Limitations

• Free tier: Lower-resolution output. Full-resolution downloads require the Pro plan ($48/yr).
• Privacy: Images are uploaded to their servers for processing. Do not use this for confidential images.
• Large text blocks: Removing text that covers 30%+ of the image degrades quality significantly — there is not enough context for the model to reconstruct from.

For images where privacy matters and you prefer client-side processing, Pixotter's background removal tool handles all processing in your browser — no upload, no server round-trip.

Method 5: Inpaint Online

Inpaint is one of the older web-based inpainting tools and still handles straightforward text removal competently. The interface is minimal: upload, mark the text with a red brush, click "Erase," and download.

Steps

1. Go to theinpaint.com.
2. Upload your image.
3. Use the marker tool to highlight the text. The red overlay shows what will be removed.
4. Click Erase. Inpaint fills the marked area.
5. Download the result. Free downloads include a small watermark; the $19.99 one-time purchase removes it.

Best for

• Quick, simple removals where the background is not highly detailed
• Users who want a one-time purchase instead of a subscription
• Situations where you need results in 30 seconds without creating an account

Limitations

• The inpainting model is older than Cleanup.pictures and produces visible artifacts on complex textures
• Free output includes a watermark (ironic for a text removal tool)
• Maximum file size is 10MB

Method 6: Mobile Apps (TouchRetouch and Snapseed)

When you need to remove text from an image on your phone — a screenshot, a photo with a date stamp, a social media image with an overlay — these two apps cover the full range.

TouchRetouch (v5.0) — iOS and Android, $2.99

TouchRetouch is purpose-built for object and text removal on mobile. The text removal mode is specifically designed for this use case.

1. Open TouchRetouch and load your image.
2. Tap Quick Repair or Text Remover (the dedicated mode for text).
3. Brush over the text. The app highlights what it will remove.
4. Tap Go. TouchRetouch inpaints the area.
5. Review and save. Undo and redo individual strokes if needed.

TouchRetouch handles multi-line text, curved text, and text on moderately complex backgrounds. The $2.99 price is a one-time purchase with no subscription and no ads. It is the best-value option on this list for mobile use.

Snapseed (v2.21) — iOS and Android, Free

Snapseed (by Google) is a free photo editor with a capable Healing tool. It is not purpose-built for text removal, but the Healing brush works on text the same way it works on any unwanted object.

1. Open the image in Snapseed.
2. Go to Tools > Healing.
3. Tap and brush over the text. Snapseed fills the area with surrounding texture.
4. Repeat for each text element.

Snapseed's Healing tool works best on smaller text. For large text blocks or text on busy backgrounds, TouchRetouch's dedicated text mode produces cleaner results. Snapseed is free with no limitations, though — if you already have it installed, it is worth trying first.

Common Use Cases

Removing watermarks from your own images

Stock photo services and camera apps sometimes add watermarks or branding to images you own. Legitimate scenarios: removing a "SAMPLE" watermark from a proof after purchasing the license, removing your own camera app's watermark, or cleaning a logo from internal company photos. For a dedicated walkthrough of this specific use case, see our guide on how to remove a watermark from an image.

For simple, repeating watermarks (like a tiled "SAMPLE" overlay), Photoshop's Content-Aware Fill handles the repetition well. For a single watermark in one corner, any method on this list works — Cleanup.pictures is the fastest if you want zero setup.

If you are adding watermarks to protect your own work, our guide on how to add a watermark to images covers positioning, opacity, and batch watermarking.

Removing text overlays from social media images

Social media platforms bake text overlays into shared images — Instagram story text, Snapchat captions, meme text. If you have the original photo and just need the overlay removed, TouchRetouch on mobile or Canva's Magic Eraser handles this quickly.

For meme-style text (white Impact font with black outline on a photo), the high contrast makes selection easy in any tool. Photoshop's Object Selection tool picks up this style of text automatically.

Cleaning up screenshots

Screenshots often contain UI text, notification bars, personal information, or debug overlays that need removal before sharing. For text at the edges (status bars, navigation), cropping is faster and cleaner than inpainting — no AI reconstruction needed.

For text in the middle of a screenshot (a chat name, an email address, a notification popup), the Heal tool in GIMP or Cleanup.pictures works well because screenshot backgrounds are typically solid colors or simple gradients — the easiest case for inpainting.

If you need to blur text instead of removing it entirely (to indicate redacted content), blurring faces or text in photos is a better approach for privacy-sensitive content.

Removing date stamps from photos

Older digital cameras and some phone camera apps embed date stamps directly into the image pixels. These are usually white or orange text in a corner, on a relatively simple background (sky, ground, wall). Every method on this list handles this well. Cleanup.pictures or TouchRetouch are the fastest — select the date text, let the AI fill in the corner, done in under a minute.

Tips for Better Text Removal Results

1. Start with the highest-resolution version of the image. Text removal works by reconstructing pixels from surrounding context. More surrounding pixels = better reconstruction. If you have the original file, use it — do not work from a compressed JPEG thumbnail.

2. Expand your selection past the text edges. Text rendering includes anti-aliased pixels — semi-transparent edge pixels that blend the text into the background. Select 2-4 pixels beyond the visible text boundary to catch these. Missing them leaves a ghostly outline.

3. Remove text before resizing or compressing. JPEG compression creates artifacts around high-contrast edges (like text on a photo). Removing text from a heavily compressed image means the inpainting model also has to reconstruct compression artifacts. Work on the cleanest source, then compress the final result or resize it afterward.

4. For repeating patterns behind text, show the AI more of the pattern. If text sits on a tiled floor or striped fabric, expand the sampling area (in Photoshop) or remove text in small sections so the tool has more pattern context per stroke.

5. Check your results at 100% zoom and on the target device. Inpainting that looks perfect at 50% zoom can show visible seams, color shifts, or texture breaks at full resolution. If the image is for mobile, check it on a phone screen.

Which Method Should You Use?

You need the best possible result and have Photoshop: Content-Aware Fill in Photoshop v26.3. Nothing else matches it on complex backgrounds.

You want free and good enough: Cleanup.pictures for browser, GIMP v2.10.38 for desktop, Snapseed v2.21 for mobile.

You are on your phone: TouchRetouch ($2.99) for the best mobile result. Snapseed (free) if you want to try without paying.

You want fast and do not want to install anything: Cleanup.pictures. Upload, brush, download.

The text is at the edge of the image: Skip inpainting entirely. Crop the image and save yourself the effort.

You need privacy (no upload): GIMP processes everything locally. For browser-based client-side processing, Pixotter's image tools run entirely in your browser with no server uploads.

Frequently Asked Questions

Can I remove text from an image for free?

Yes. GIMP v2.10.38 (desktop), Snapseed v2.21 (mobile), and Cleanup.pictures (browser, limited free tier) all handle text removal without payment. For simple backgrounds, these free tools produce results comparable to Photoshop.

Does removing text reduce image quality?

The inpainted area is reconstructed — it will never be identical to what was originally behind the text. On simple backgrounds, the reconstruction is seamless. On complex textures, there may be subtle differences visible at high zoom. Working from the highest-resolution source minimizes quality loss.

How do I remove text from a screenshot without leaving artifacts?

For text at the edges, crop it out — this is faster and leaves no artifacts. For text in the middle, use the Heal tool in GIMP or Cleanup.pictures. Screenshot backgrounds are typically solid colors or gradients, which are the easiest case for inpainting and rarely produce visible artifacts.

Can I remove watermarks from images I downloaded?

Only if you have the legal right to do so — for example, after purchasing a license for a stock photo that was previewed with a watermark. Removing watermarks from images you do not own or license is copyright infringement under most jurisdictions.

What is the difference between removing text and extracting text?

Removing text erases the text from the image and fills in the background. Extracting text reads the text and copies it as editable characters (OCR). They are opposite operations — one destroys the text, the other preserves it.

Does Photoshop's Content-Aware Fill use AI?

Yes. Starting with version 26.x, Content-Aware Fill uses Adobe's Firefly generative model for reconstruction. Earlier versions used a non-neural sampling algorithm. The generative version produces significantly better results on complex textures and patterns.

Can I remove text from a video frame?

These tools work on individual images. To remove text from video, export the frame as a PNG, remove the text with any method above, and replace the frame. For persistent text (watermarks, subtitles) across many frames, video-specific tools like DaVinci Resolve 19's object removal or After Effects' Content-Aware Fill for video are more practical.

What is the best free mobile app for removing text from photos?

Snapseed v2.21 (by Google) is free with no limitations and handles simple text removal well. For better accuracy and a dedicated text removal mode, TouchRetouch v5.0 costs $2.99 as a one-time purchase — no subscription, no ads.