
Photo Cleanup API by Atlas Cloud
Restore single images by reducing dust, scratches, noise, and repeated micro-texture artifacts without changing the image content.
Photo Cleanup
Photo Cleanup restores single images by reducing dust, scratches, noise, and repeated micro-texture artifacts. It is designed for post-processing generated or uploaded images that need a cleaner surface without changing the image content.
Highlights
- Image cleanup chain: Runs dust cleanup followed by denoise cleanup for a balanced restoration pass.
- Dust and scratch reduction: Softens small specks, scratch-like marks, and dirty surface artifacts.
- Noise and micro-texture reduction: Reduces repeated fine texture, scale-like artifacts, and noisy high-frequency edges.
- Content-preserving output: Keeps the original composition, subject, framing, and color intent.
- Single-image workflow: Optimized for still images. Video cleanup is not supported by this model.
Parameters
| Parameter | Required | Description |
|---|---|---|
model | Yes | Use atlascloud/photo-cleanup. |
image | Yes | Input image URL. |
output_format | No | Output image format: png, jpeg, webp, or jpg. Default: png. |
Limits
Photo Cleanup is an image-to-image model. It does not upscale images, does not process video input, and does not expose the internal single-stage cleanup runners as public parameters.
Use input images that are accessible by URL. Very large inputs may be rejected by the backend before processing.
Example Request
{ "model": "atlascloud/photo-cleanup", "image": "https://atlas-test-input.oss-us-west-1.aliyuncs.com/1024_square.png", "output_format": "png" }
Best For
- Cleaning generated images with visible repeated micro-texture or scale-like artifacts.
- Reducing dust, fine scratches, and small specks in still images.
- Softening noisy water, metal, foliage, glass, or reflective texture regions.
- Preparing images for visual review when the source image is already close to usable.
Pricing
Photo Cleanup uses flat per-image billing.
| Unit | Price |
|---|---|
| Per image | $0.0200 |
Notes
- Cleanup reduces artifacts; it does not guarantee complete artifact removal.
- Strong cleanup can increase edge artifacts or soften real detail.
- Severe generation artifacts are better handled with rerolling or localized regeneration.
- This model does not upscale images and does not process video input.



















