Remove Sticker from Image
Erase stickers, emoji, chat bubbles, and overlay graphics from any screenshot or photo. SDXL generative fill rebuilds the content underneath. Free, no signup, 3 removes a day.
Before / After
Erase unwanted objects and clutter
Why generative fill beats simple inpainting here
Stickers and emoji are unusually hard for traditional inpainting because of what is typically underneath them. A watermark sits on photographic background — sky, foliage, fabric, a predictable texture. A sticker, in contrast, often sits on a face, on chat UI, on text, or on a complex composition where context from neighbouring pixels is misleading. Reconstructive inpainters (LaMa, Content-Aware Fill) blur the masked area into the closest matching surroundings; the result on a face is usually an unsettling smear.
SDXL inpainting takes a different approach. It generates new pixels from its learned model of what photos and screenshots look like, conditioned on a prompt that says "clean original content, restored background". When the sticker is over a face the model generates a plausible face; when it's over a chat bubble the model generates plausible UI chrome. The result is not the actual content that was underneath, but it looks like it belongs.
The tradeoff: SDXL runs its VAE over the whole image, so pixels outside the mask drift by a small amount. For social media cleanup this is almost never noticeable. For strict pixel-level preservation (stock photo watermarks, product photography) use the LaMa-based tools instead. This page is the right tool for Instagram stories, Snapchat exports, WhatsApp screenshots, and meme pictures — the SDXL losses don't matter and its generative strengths do.
How to remove stickers in 3 steps
- 1
Upload the screenshot or photo
Drag any JPG or PNG into the upload card — chat screenshots, Instagram stories, TikTok exports, meme pictures, anything with stickers or emoji pasted on top. Images larger than 1536 px on the longest edge are automatically downscaled client-side before upload. Aspect ratios are preserved; SDXL does not lock the output to a square.
- 2
Brush every sticker, emoji, and overlay
Paint over each element you want gone — smiley-face stickers, heart reactions, chat bubbles, platform watermarks, filter effects, censor bars. For stickers with hard edges (most platform-generated emoji) add 4–6 pixels of margin beyond the sticker border so SDXL has clean context to rebuild. For layered overlays (filter frames that wrap the whole image), paint each element separately for best results.
- 3
Download the clean version
Tap Remove sticker. SDXL inpainting takes roughly 20–30 seconds to redraw the masked regions with plausible content — skin tones, background textures, text behind the sticker, whatever was there originally. Hit Download for a PNG at your upload's dimensions. Sharing the clean image back to social without the original platform's sticker is now possible, though do respect each platform's terms of service.
Where this tool earns its keep
Instagram and Snapchat stickers
Location tags, GIF stickers, countdown timers, music stickers, AR filter frames. SDXL handles the round corners and animated-looking shapes that LaMa sometimes over-smooths.
Emoji reactions pasted on faces
A coworker's shared screenshot with a 😂 slapped over someone's face — paint the emoji and SDXL rebuilds the face underneath. Expect good but not photographic results; details like exact eye expression are guessed.
Chat bubble balloons and message overlays
WhatsApp, iMessage, Telegram, WeChat bubble shapes covering the background of a screenshot. SDXL reconstructs the app UI or wallpaper behind the bubble.
Platform watermarks and branding
TikTok usernames, Snapchat time stamps, Instagram 'remixed from' labels — anything a social app stamps onto a shared image.
Censor bars and privacy mosaics
Black bars across eyes or sensitive text, pixelation mosaics over license plates or name tags. SDXL can sometimes rebuild underneath, though heavily pixelated regions have less context to work with.
Meme text and overlay typography
Classic meme-format bold-white text, TikTok-style subtitle overlays, 'POV' labels, reaction captions. The text is painted over just like any other sticker.
Tips for the cleanest results
Paint the whole sticker including drop shadows
Many stickers have a subtle drop shadow beneath them. A tight mask that only covers the opaque pixels leaves a faint halo. Zoom in and paint 3–5 pixels past the sticker's visible edge to catch the shadow.
For animated GIF stickers, upload a still frame
GIFs are a sequence of frames. Upload the still frame you care about cleaning; SDXL works on a single image at a time. To clean a full GIF you would need to process each frame individually and re-encode.
Very large stickers covering faces will be approximations
If a heart emoji covers most of a person's face, SDXL has very little to work with. The result will be a plausible-looking face but NOT the original person's actual features. Use this for aesthetic recovery, not identity recovery.
Recover platform UI text behind bubbles
SDXL is surprisingly good at rebuilding recognisable iOS / Android UI text when a chat bubble covers it. The exact message text will be invented (SDXL has no way to know what the original said), but the UI chrome is usually restored well.
The honest limits
Generative fill means pixel drift everywhere
SDXL redraws the whole image — unmasked regions are visually close to the original but not byte-identical. For strict preservation (watermark removal on detailed product photos), use the LaMa-based Watermark Remover instead.
Faces behind stickers will look 'like a face' not THE face
When a sticker covers a face, SDXL invents facial features from context. The reconstructed face is anatomically plausible but not the actual person's likeness. This is why aesthetic cleanup works but identity recovery does not.
Heavily pixelated censors cannot be reliably 'decensored'
Aggressive mosaic or black-bar censoring destroys information. SDXL will invent plausible pixels underneath, but those are fabrications, not recovered originals. Do not use these results to identify censored people or read blurred text with any confidence.
Questions about removing stickers and emoji
Other removers on MagicRemover
Free AI Object Remover
The homepage pill — start here for simple removals. LaMa-based, fast, and preserves surroundings.
Remove Watermark from Photo
LaMa-based watermark remover. Byte-preserving — best when the rest of the image must stay exactly as captured.
Remove People from Photo
SDXL-based background people remover. Handles crowds and passers-by with full scene rebuild.