Remove unwanted objects from photos using AI inpainting. Paint and erase.
Remove unwanted objects, people, or blemishes from photos using AI inpainting. Simply paint over the area to remove and the AI fills it in naturally.
Upload your file using the tool above.
Adjust any settings to your preference.
Click the process button and wait for results.
Download your output using the download button.
Paint slightly beyond the object’s edges, including its shadow and any reflection — leaving those behind is the most common reason a result looks wrong. Objects against simple, repetitive backgrounds like sky, grass or a plain wall fill in almost invisibly, because there is a clear pattern to continue.
The model invents what it thinks belongs there, so it struggles when the background is complex or when the object sits over something with strong structure — a face, text, a straight architectural line. Removing a large object from a busy scene gives it too much to guess at.
Work one at a time for the best results. Each removal gives the model a cleaner image to reason about for the next, whereas masking half the photo at once forces it to invent a large connected region. It is slower but noticeably more convincing.
Yes, AI Object Remover is completely free. No signup, no account, and no watermark on outputs. A Pro tier is available for 100 AI ops/day and larger file sizes.
AI Object Remover sends your file to our AI provider over an encrypted connection so the model can process it. It is deleted after processing and is not used to train models. If a file is confidential, use one of our in-browser tools instead, which never upload anything.
Drop a photo — paint over objects to remove them
PNG, JPG, WebP — up to 10 MB · free · no sign-up required
This tool is powered by a state-of-the-art large-mask inpainting AI model that understands repeating patterns (bricks, fences, grass, fabric) and fills masked regions seamlessly. It runs on an NVIDIA T4 GPU — serverless, so there is no idle cost and no long wait after the first warm-up request.
Quick mode caps the input at 512 px (~3–5 s) — fast and accurate for clean, uniform backgrounds like sky, walls, and floors.
Quality mode caps the input at 1024 px (~6–10 s) — processes at twice the resolution for sharper fills on complex textures and irregular objects.
Best results tips:
Supported inputs: JPG, PNG, WebP · Max 10 MB · Tools: Brush, Eraser, Freehand Lasso · Undo: Ctrl+Z · Output: PNG · AI on a T4 GPU (Quick = 512 px, Quality = 1024 px)