Yes. OSS models are Apache-2.0. The hosted API is freemium: 50 free credits, then pay-as-you-go. Open source: commercial use, modification, redistribution permitted. Hosted API: newer weights, faster inference, autoscaling. Python package: local OSS inference, or add api_key for cloud. Pricing: credits per image. No subscription lock-in.

Open-source background removal and API

Q: Do you store or train on my images?

No. Images are processed in RAM and discarded after the response. Storage: no disks, thumbnails, caches, or backups. Region: Frankfurt, Germany (EU). Logs: metadata only (timestamp, endpoint, duration, status, size, key hash). Training: customer uploads are never used for model training.

Q: Do you use customer data for training?

No. Customer uploads are never used to train or fine-tune models.

Self-host the OSS model for full privacy, or use the API for faster inference on harder images. Compare sample results on hair, fur, and glass.

Precision Cutouts
Hair, Fur, Glass, Motion
Open source
Apache-2.0 · pip install · full model weights
Self-hosting
Docker, local Python, zero outbound image data
Production API
0.05€/image · p95 ≈ 800 ms · Frankfurt, DE · AWS Inferentia

Get API key API docs Sample results

Self-host: github.com/withoutbg/withoutbg · Try the demo

Fine hair strands, backlight, motion

Portrait with windblown curly hair and dramatic backlight

Same portrait with background removed, hair edges preserved

OriginalCutout

Alpha mattePredicted segmentation mask

Edge regionat 2.8× zoom · 0%, 22%

Open-source background removal and API

Self-host the OSS model for full privacy, or use the API for faster inference on harder images. Compare sample results on hair, fur, and glass.

Precision Cutouts
Hair, Fur, Glass, Motion
Open source
Apache-2.0 · pip install · full model weights
Self-hosting
Docker, local Python, zero outbound image data
Production API
0.05€/image · p95 ≈ 800 ms · Frankfurt, DE · AWS Inferentia

Get API key API docs Sample results

Self-host: github.com/withoutbg/withoutbg · Try the demo

Quick start

Run instantly with Docker

docker run -p 80:80 withoutbg/app:latest

Python package

pip install withoutbg

View on GitHub1,089 stars

Integration paths

Cloud API, Docker self-host, or local Python. Same API schema for local and cloud paths.

Cloud API

Pro model

Managed inference

Get API key

REST API hosted in Frankfurt, DE. p95 ≈ 800 ms server-side. Strong on hair, fur, and fine edges. See API results for examples.

Inferentia · p95 ≈ 800 ms · EU (Frankfurt)

Docker

Self-hosted

Self-hosted web UI

Docker Hub

One-command deployment with the OSS model. Images stay on your network; no outbound uploads required.

Copy and run

docker run -p 80:80 withoutbg/app:latest

Python SDK

Apache-2.0

PyPI · local or API

PyPI

Run the OSS model locally, or pass api_key to route through the cloud API. Same request schema for both paths.

Copy and run

pip install withoutbg

Integrations

In development

Native plugins for design and editing tools.

Blender

Figma

GIMP

macOS

Windows

API pricing

Pay-as-you-go credits. OSS models are free under Apache-2.0. API from €0.05/image at volume.

Package	Credits	Price	Per credit	Validity	Savings
Free trial	50	€0	€0.00	30 days
Package XS	100	€10	€0.10	30 days
Package S	500	€40	€0.08	30 days	20% off
Package M	1,000	€70	€0.07	90 days	30% off
Package L	5,000	€300	€0.06	180 days	40% off
Package XL	10,000	€500	€0.05	360 days	50% off

Specs and limitations

Specifications, security, training data, and known limitations for the API and open-source models.

Background removal models

Cloud API: Hosted background removal with alpha matte output. p95 ≈ 800 ms server-side in Frankfurt, DE. See sample results.
OSS model: Apache-2.0 weights for local inference via Python or Docker. See OSS sample results.
Both return PNG cutouts and optional alpha mattes for compositing.

Security and privacy

OSS: Run locally; no API calls required. Full control over image data.
API: TLS in transit; in-memory processing only; no disk writes of customer images.
Logging: Timestamp, endpoint, duration, status, request size, API key hash. No image bytes or perceptual hashes.
Analytics: No cookies on the web UI. Ahrefs (privacy-friendly). In-house CAPTCHA for abuse prevention.

Training data

No customer training: Uploads are never used to train models. Buffers zeroed after response.
Public photos: Unsplash/Pexels under permissive licenses; we created alpha mattes. withoutBG100 dataset
Licensed sets: Purchased image sets with explicit derivative rights.
Synthetic renders: Randomized lighting/camera with ground-truth mattes from the render pipeline.
Studio captures: Hair, translucency, shadows. Dataset guide
Scale: ~60K image/matte pairs after QA (2025-10-01), expanding.

Limitations

Transparent materials: Glass, water, and sheer fabric remain inconsistent due to complex light transport.
Subjective boundaries: When foreground vs. background is ambiguous, multiple valid cutouts exist. Future release will bias toward nearest-camera subjects.

FAQ

Is it really free?

Yes. OSS models are Apache-2.0. The hosted API is freemium: 50 free credits, then pay-as-you-go.

Open source: Commercial use, modification, redistribution permitted.
Hosted API: Newer weights, faster inference, autoscaling.
Python package: Local OSS inference, or add api_key for cloud.
Pricing: Credits per image. No subscription lock-in.

How fast is it end-to-end?

End-to-end depends on upload size and network RTT. We publish server-side processing only: p95 ≈ 800 ms in Frankfurt, DE. Total time = upload + TLS + queue + ~800 ms + download.

Do you store or train on my images?

No. Images are processed in RAM and discarded after the response.

Storage: No disks, thumbnails, caches, or backups.
Region: Frankfurt, Germany (EU).
Logs: Metadata only (timestamp, endpoint, duration, status, size, key hash).
Training: Customer uploads are never used for model training.

Do you use customer data for training?

No. Customer uploads are never used to train or fine-tune models.

Supported by