Bulk photo work — product cards, site photos, shoot archives — is manual labour in an editor and, in cloud services, runs into per-image billing plus uploading client material to a third party. Open models cover the individual operations: strip the background, upscale and clean up an old photo, remove an unwanted object or caption, restore a face on a blurry shot. All of it installs on your own server and can be wired into a pipeline that processes a thousand cards without a designer, while the originals stay with you. What does not appear is a general-purpose editor: each model does one thing and chaining them into a process is your work. On hard edges — hair, fur, transparent and reflective surfaces — results are uneven, and part of the output still goes to a designer for cleanup.
An open MIT-licensed model for precise object segmentation and background removal. RMBG-2.0 is built on it. Versions for 2K and for hair and semi-transparent edges.
A background removal model focused on difficult edges: hair, fur, fine details. The open version is MIT-licensed and can process video.
The classic for upscaling photos 2–4x while cleaning noise and compression artifacts. Lightweight, runs even on a CPU. Versions for drawings and anime.
Removes unwanted objects, text and watermarks from photos with clean background fill. Lightweight and fast; still the standard for this task.
Proven face restoration for old and compressed photos, with a commercial-friendly license. Often paired with Real-ESRGAN; the most used versions are 1.3 and 1.4.
These five are listed as commercially usable (MIT, Apache 2.0, BSD-3). The popular BRIA RMBG, CodeFormer and SUPIR that do similar jobs are non-commercial and cannot go into a business pipeline without a separate paid license.
An open model can run on your own server: data stays in-house, there is no per-request fee, and the model can be fine-tuned on your documents.
