Privacy Policy.

At BG Remover, we believe in full technical transparency. Our application uses a hybrid dual-engine architecture:

• Client-Side WebAssembly (Default / Fallback Engine): By default, neural network inference runs 100% locally inside your web browser via WebAssembly (WASM) and WebGL. When running in WASM mode, zero image bytes ever leave your local computer.
• Optional Server AI Microservice: When available, a high-precision server AI endpoint (utilizing SOTA segmentation models like BiRefNet) may be invoked to handle ultra-complex image cutouts.

Whether processed via client-side WebAssembly or the optional server API:

• No Permanent Storage: Images sent to the optional server microservice are processed strictly in ephemeral RAM and immediately discarded upon completion.
• No File Logging: We do not log original photo uploads, output cutout files, or facial features to disk.
• No AI Model Training: We do NOT use your uploaded photos to train or fine-tune public AI models.
• No Third-Party Data Sales: Your images and personal visual assets are never sold, shared, or distributed.

We do not use invasive cross-site tracking cookies. We utilize standard browser local storage solely to save user interface preferences (such as aspect ratio crop defaults or studio background selections). Anonymous pageview analytics are collected via Privacy-Friendly Vercel Analytics.

If you have technical questions regarding our WebAssembly architecture or privacy implementation, please contact our team at privacy@bgremover.com.