The World's Largest Public Domain Media Search Engine

AI-Driven Search Has Arrived

01/25/22 19:31:00

Since October 2021, our library of public domain media has grown by 2 Million images reaching the 24 Million documents milestone!

All media files we add are diversified by topics and annotated by our dedicated team. About 20% of media files we add are already sorted out into human-curated collections.

Now, we announce a transition from a tag-based search engine to a much more nuanced natural language neural-networks-based one. Our new AI works best with human-written content, and sometimes it looks as if it "understands" the meaning of phrases, not just keywords, delivering amazingly relevant images that no other method can deliver. It is also important that AI is capable to understand multi-language metadata and find matching images that come from sources in languages other than the search request language. 

We also work hard to rewrite and enhance the titles and descriptions of millions of images to make them more human-friendly and readable by the AI engine we built. This already resulted in search improvement for many popular requests. We will keep working on this until delivers more meaningful results for requests like "Portraits of those who signed the Declaration of Independence", "Roman battles", "Historic sites in Bavaria" and so on.

In addition to 20+ Million English titles and descriptions, we also generated about 35 Million Russian and German translations - all towards making the largest and the best public domain search engine in the world.

Find all recent updates here: GetArchive Progress Report