rembg
Remove image backgrounds locally through a Python command, library or self-hosted HTTP service.
- Starting price
- Free local Python software
- Free access
- Free MIT package
- Billing unit
- Local compute or your hosted server
- Platform
- Python CLI, library and HTTP server
- Verified
- 20 Oct 2026
Overview
What you provide
- Provide an image file, folder or byte stream and choose a suitable segmentation model.
What you get
- rembg returns foreground cutouts with transparent backgrounds through local processing interfaces.
Choosing this tool
rembg fits repeatable cutout jobs when files should stay on your own machine or server. The CLI is useful for scripted folders, while the library or HTTP mode fits an existing application. Clipdrop is simpler for occasional browser editing; backgroundremover offers another local video-and-image removal pipeline with its own dependencies.
A decision worked through
A folder containing 500 images does not need a 500-credit service pack when processed locally with rembg. First test hair and translucent edges on ten representative files; any reruns consume machine time and disk rather than a published hosted image allowance.
rembg alternatives for this task
For rembg, compare Clipdrop, backgroundremover. Match the input, delivery format and operating responsibilities before comparing price.
| Product | Price / unit | Free allowance | Input → output | Key limit |
|---|---|---|---|---|
| rembgRead profileOfficial source ↗ | MIT package: no image fee | Local files; no hosted credit quota | Photos become transparent cutouts. | Choose compatible ONNX/backend packages. |
| ClipdropRead profileOfficial source ↗ | CNY 119/mo Pro | 20 background removals/24h | Images become background-free cutouts. | Pro adds high-resolution removal; API billed separately. |
| backgroundremoverOfficial source ↗ | MIT software; local resources extra | Local image/video processing | Images/videos become cutout media. | Python, ffmpeg and model setup. |
Setup & workflow
- Install the documented CPU or GPU package variant in a compatible Python environment.
- Choose an explicitly reviewed local model, such as u2net, and inspect one output alpha mask.
- Process the folder or integrate the library only after the sample edges pass review.
Requirements & installation
- Use Python >=3.11,<3.14 and pip install "rembg[cpu,cli]" for CPU CLI use; download weights before offline processing.
Limits & review
- The current default bria-rmbg model requires a paid agreement for commercial use; that amount is not published here. Choose and review another model, such as u2net, deliberately. Cloud backends such as withoutbg upload images and have separate service bills; installing rembg alone does not make every backend local or free.
Your part
- Check edge masks and confirm that product details and shadows are retained as intended.
Sources & verification5
This profile is based on official sources, not a hands-on product test.
Vendor descriptions and demos document advertised features. Editorial guidance is based on these sources.
- Official CPU/GPU installation and model optionsgithub.com
- Official u2net source and licensegithub.com
- rembg official referencegithub.com
- Clipdrop official referenceclipdrop.co
- backgroundremover official referencegithub.com
