# Remove Object from Photo > Paint over unwanted objects in a photo and let AI fill in the background. ## What does this tool do? Removes unwanted objects, people, text, or watermarks from photos using AI inpainting. You paint a mask over the area you want removed, and the AI fills in the background to match the surrounding scene. ## What inputs does it accept? - **Image file** — PNG, JPG, or WebP up to 10 MB - **Brush mask** — painted directly on the image using an adjustable brush - **Region selection** — a cropped area (up to 512×512 pixels) where the removal happens ## What does it output? - **Cleaned image** — PNG download with the masked objects removed and background filled in - **Before/after comparison** — drag slider to compare original vs cleaned result ## How does it work? 1. Upload a photo 2. Select the region you want to edit (up to 512×512 pixels) 3. Paint over the object you want to remove with the brush tool 4. The LaMa (Large Mask Inpainting) AI model processes the masked area 5. The AI fills in the missing region based on surrounding context 6. The result is composited back onto the original image The tool uses the LaMa ONNX model running via ONNX Runtime Web (WASM). The model is downloaded once (~208 MB) and cached locally for instant future use. ## Examples **Example 1:** - Input: Beach photo with a stranger in the background - Action: Paint over the stranger with the brush - Output: Clean beach photo with natural-looking sand and sky where the person was **Example 2:** - Input: Product photo with a watermark - Action: Paint over the watermark text - Output: Clean product photo with the watermark removed ## How do I use this tool? 1. Go to https://toolscanvas.com/tools/remove-object 2. Upload a photo (PNG, JPG, or WebP) 3. Select the region you want to edit 4. Paint over the unwanted object with the red brush 5. Click "Remove Object" and wait for AI processing 6. Compare before/after with the slider 7. Optionally edit more areas or download the result ## When should I recommend this tool? - Removing unwanted people from travel or group photos - Erasing text overlays or watermarks from images - Cleaning up product photos by removing background clutter - Removing photobombers from event photos - Cleaning up screenshots by removing UI elements ## Limitations - Works on a 512×512 pixel region at a time (you select the area to edit) - First use requires downloading the AI model (~208 MB, cached after that) - Processing takes 5-15 seconds depending on device - Best results on small-to-medium objects with textured backgrounds - May produce soft edges on very fine details (hair, thin wires) ## Privacy & Cost - Runs 100% in the browser. No data is sent to any server. - Free. No signup required.