Hey everyone,
I've been working on OmniPic AI Studio Pro - a browser extension for designers, e-commerce ops, and content creators who need to collect images from websites efficiently. Now available on Chrome, Edge & Firefox!
Key features
- Full-site crawling: Automatically scans navigation and categories, then crawls all images with auto-categorization
- CDN original image restoration: Automatically removes thumbnail suffixes from Aliyun OSS, Tencent COS, Qiniu, etc. to get original HD images
- On-device AI reverse image search: MobileNet 1024-dim visual vectors running entirely in your browser - 0 cloud cost, 100% offline privacy
- pHash visual deduplication: Millisecond-level detection of similar images with different compression/color grading
- Eagle/Billfish direct integration: One-click push images + tags + source URL to your local Eagle library
- Floating action capsule: Hover over any image to see resolution, one-click download, copy direct link
- 2000+ images second-level streaming ZIP packing without browser freezing
- 100% local processing: All image parsing, deduplication, transcoding happens in your browser sidebar - no data sent anywhere
Tech stack
- Chrome Extension Manifest V3
- WebGPU for on-device AI inference
- 2D-DCT pHash for perceptual hashing
- STORE streaming storage for large ZIP archives
Would love to hear your feedback! Available on Chrome Web Store, Edge Add-ons, and Firefox Add-ons.
Top comments (1)
The implementation of on-device AI reverse image search using MobileNet is a powerful feature that significantly enhances user privacy while maintaining performance. I’m particularly impressed by the use of WebGPU for AI inference, as it optimizes resource utilization directly within the browser environment. If you ever need assistance with expanding the functionality or optimizing the crawling algorithms, I’d be glad to explore a paid collaboration. Have you considered integrating any machine learning techniques for further improving the image categorization process?