BackgroundRemover
CLI tool and HTTP API for removing backgrounds from images and videos using AI models and FFmpeg.
Extract backgrounds from video files and images locally using PyTorch models and FFmpeg integration.
Backgroundremover is an open-source command-line tool and local HTTP server that isolates foreground subjects in photos and video clips. Built on PyTorch and FFmpeg, it provides a fast, scriptable alternative to cloud rotoscoping services for video editors and pipeline developers.
What it is
The tool uses pre-trained U2Net neural network models to generate binary alpha masks and transparent video files in formats like MOV. You can process individual media files, run batch operations across entire folders, fine-tune edge quality using alpha matting settings, or launch it as a local REST API endpoint.
Who it's for
It suits video editors, developers, and visual effects technical directors who need automated keying or background removal via terminal commands and build scripts. It is not suited for users who require a graphical interface or timeline-based video editor.
Key features
Transparent video export
Processes video files using FFmpeg to output MOV clips with alpha channels or transparent backgrounds.
Configurable alpha matting
Fine-tune edge erosion, foreground thresholds, and background sensitivity for detailed cutouts.
Multiple neural models
Switch between standard U2Net, lightweight U2NetP, or specialized human segmentation models depending on speed and subject requirements.
Batch and pipe processing
Process entire image and video folders or stream images through Unix standard input and output pipelines.
Local HTTP API server
Launch an internal web server to accept REST API POST requests and image URLs for automated workflows.
Up and running in 3 steps
Our honest verdict
What's great
- Exports transparent MOV video files directly via FFmpeg integration.
- Includes alpha matting parameters to adjust edge sharpness and erosion.
- Runs locally without sending media to external third-party servers.
- Supports standard input and output Unix pipelines for automated scripting.
Where it falls short
- Requires setting up system dependencies like FFmpeg 4.4+ and Python development packages.
- Video processing demands significant shared memory and GPU hardware for fast frame rates.
- Lacks a graphical user interface, requiring all operations to run through the command line or API.
Pricing & license
Backgroundremover is free to self-host and run under its open-source licence. The creator operates a hosted web application called BackgroundRemoverAI.com, but running the CLI or Docker container locally incurs no usage fees beyond your compute hardware costs.
Open remote AI-training roles
876 open remote AI-training roles, updated November 15, 2026; most pay $20–$170 an hour. Zekai lists them free; the application goes to the employer’s platform.
Alternatives
Use it, or have us set it up
FAQ
Can I process an entire directory of files at once?
Yes, use the --input-folder and --output-folder flags to batch process supported image and video formats.
Does backgroundremover require a GPU?
No, it automatically falls back to CPU execution, though processing speeds will be significantly slower without CUDA acceleration.
Can I host backgroundremover as a microservice?
Yes, running backgroundremover-server starts an HTTP REST API server on your chosen host port.
