Intuitive Drag and Drop Interface
Compendia makes organizing your photos feel simple right from the start. The drag and drop tagging interface is easy to understand and fun to use, so you can just dive in to organizing your library without a steep learning curve. Adding tags is quick and natural. You can work incrementally and save or return later, or work through smaller subsets of your library one at a time.
Instant Search
Compendia gives you fast, flexible search across your entire library. You can find photos by tag, capture date, file location, or file name, and results appear as soon as you start typing. It feels immediate and predictable, and it helps you locate files even the largest collection.
Lightweight but Powerful
In the spirit of modern applications, compendia is small and lightweight – but it’s engineered with a two-layer cache that allows you to navigate smoothly through even tens of thousands of image previews.
Modern Automation
The application can help you with modern features that automatically search your library, but without creepy online services that require sending your data to unknown places on the internet.
Private, Fully Offline Facial Recognition
Compendia’s facial recognition works by enabling you to tag a few people that appear in your library and then it will find other matches automatically. It’s a very simple concept for fully offline face recognition – you make the initial set of faces to recognize, and the application just extends that by finding matching faces within your own collection.
Image Similarity and Duplicate Detection
You may have true duplicate images stored, maybe from repeated phone or camera dumps. Compendia has an image similarity feature that will pair up not just identical files but also similar images, and put them in sets for you to review.
Auto Tag by Capture Date
Image files from phones and cameras store valuable metadata in an internal format called EXIF – but ordinarily that data is locked away inside the files. Compendia will extract all of it as part of the normal process of reading the library, and make it visible to you, stored as separate, more accessible metadata.
A simple but valuable data point is the image capture date. Using that, compendia can automatically label your library by the year and month pictures were taken, and quickly and accurately take you back to sets of files from a particular time period. For files with timestamps in the name, as is common with modern phones, compendia can also automatically mine the file names for dates.
Your Files Your Way
We built compendia with the philosophy that you want to keep your image collection. It works fully locally, over files you can store in any directory/folder structure you like. It will not move your image files, and it doesn’t upload them to a service. The metadata you create to organize the files is stored in human-readable files alongside each picture or video, not locked in a database or proprietary format.
When we offer automated ‘ai’ processing, we have kept it local, using local models.
Modular and Portable
Large libraries of images are naturally spread across multiple folders. Part of compendia’s core architecture is that each folder and all its content is a modular and self-contained collection. Those collections can be worked on separately or seamlessly integrated into a larger library just by pointing compendia at a parent or child folder, at any time.
Want to reorganize the folders’ physical structure on your hard drive? No problem. Since each folder tagged by compendia is an independent module, you can reorganize them at will, and the application will still work.
New computer? Just move your library and it will all still work. Changing from Windows to Mac or from Linux to Windows? No problem: compendia is fully cross platform with no platform-specific files.
A further benefit of this design: if you have a cloud backup solution, it will automatically back up the metadata created by compendia, since it is 100% files in the same location with your images. No database to worry about getting lost.
