Cullit vs ShotSieve
Cullit
cullit.appShotSieve
github.comSide by side
| Figure | Cullit | ShotSieve |
|---|---|---|
| Price from | — | Free |
| Free plan | — | Yes |
| Free trial | Yes | — |
| Open source | — | — |
| Windows | — | Yes |
| macOS | Yes | Yes |
| Linux | — | Yes |
| Plans published | 2 | 1 |
| Spec score · AI Photo Culling Software | 8.0 (#11) | 7.3 (#26) |
| Integrations | Cullit exports picks to Lightroom, Capture One, or a folder, and writes star ratings and color labels to XMP sidecars read by Lightroom Classic. | No third-party service integrations are listed; model assets are obtained through the learned-IQA backend and stored in configured upstream caches. |
| Intended users | The product is presented for wedding, sports, and portrait photographers. | The project is aimed at photographers and hobbyists who want help narrowing large photo folders while keeping the workflow local. |
| What it does | Cullit is a photo-culling app for photographers, with on-device AI assistance for reviewing shoots. | ShotSieve analyzes local photo folders with learned image-quality models and helps photographers review keep or reject decisions. |
Plans
In their own words
ShotSieve
ShotSieve is a free, open-source application for reviewing local photo folders with learned image-quality models. It helps photographers make keep-or-reject decisions while leaving the photo library on the user’s machine. TOPIQ is the default model, with CLIPIQA and Q-ReAlign Mini also available; users can compare models on the same library before a larger culling pass. Review tools filter and sort by score, format, resolution, file size, and metadata completeness, and support batch actions. Listed capabilities include RAW photo support, blur detection, and AI quality scoring. The review interface runs locally on loopback at 127.0.0.1:8765 by default. Releases are listed for Windows, Linux, and Apple Silicon macOS. CPU is the fallback when compatible accelerator hardware is unavailable. Model weights must be downloaded separately. The application uses AGPLv3 or later; PyIQA and model checkpoints have separate terms.