Photo to cutout
Remove the background from an image and download PNG
Upload a product photo, portrait, or object image. The image model isolates the main subject and returns a transparent result that you can inspect and download as a PNG. Choose a source image where the subject is visible and reasonably separated from the background.
Output
Transparent preview
Your image will appear here
The checkerboard reveals transparency. Download is always encoded as PNG.
How image background removal works
The uploaded image and a transparency instruction are sent to the configured image-to-image model. The model decides which pixels belong to the subject and which should become transparent. Unlike a simple chroma key, it can preserve changes in lighting, color, texture, and shape across the subject.
The returned file is checked for a valid PNG alpha channel before Pellucen shows the result. The browser then creates a checkerboard preview and a PNG download so the transparent pixels remain transparent.
Source quality matters. Sharp focus, even lighting, and clear separation between subject and background usually produce cleaner edges. Motion blur, extreme noise, a background with the same color as the subject, and objects that extend beyond the frame make the task harder for any model.
For a generated subject instead of an uploaded photo, use the Qwen-Image-2.1 image generator or start with the prompt guide.
Difficult edges to inspect before publishing
Hair and fur
Look for a hard halo where partial alpha should show fine strands. Zoom to 100% and inspect the edge over both light and dark colors.
Glass and smoke
Transparent materials contain low-opacity pixels. Check that the subject did not become a solid silhouette with all internal transparency removed.
Products and shadows
Decide whether contact shadows belong to the product asset. A cast shadow requires a surface, while a soft internal reflection may still be useful.
If you need a more controlled result, read the transparent PNG maker guide and inspect the alpha channel rather than judging transparency only from the browser preview.
Best results and practical limits
Upload a complete subject whenever possible. Cropped limbs, hidden product edges, and objects outside the frame force the model to guess what should exist. If the subject touches the image boundary, leave enough room for the model to understand the full silhouette.
Start with a clear instruction such as “isolate the main product and preserve its edge detail on a fully transparent background.” Add object-specific details only when needed. If the result removes something important, describe what must remain. If it keeps part of the background, name the object that should be removed.
This is a fast first-pass tool, not a replacement for expert retouching. Complex composites, product packaging with tiny text, and images containing many overlapping transparent objects may require manual mask cleanup after download. Return to the transparent background generator whenever you need another result.
Which image files can I upload?
Pellucen accepts JPG, PNG, and WebP files up to 10 MB. The server checks the actual file signature instead of trusting the extension or browser MIME value alone.
Will hair and fur be preserved perfectly?
Hair and fur are difficult because their visible pixels cover many partial alpha levels. A native RGBA model can preserve more soft-edge information than a simple color threshold, but fine strands may still require manual refinement.
What happens to my uploaded image?
The image is sent through the server to the configured image model for processing. Generated results expire from the current result store after 24 hours. Do not upload confidential or regulated images unless the site’s data policy is appropriate for that use.