Public Domain Image Search

Name: Public Domain Image Search
Availability: InStock
Author: jlauf

by jlauf·Feb 22, 2026·2 points·0 comments

Visit Project View on HN

AI Analysis

●●SolidSolve My ProblemDark Horse

Vector search inside images beats caption/title matching for finding obscure public domain art.

Strengths

•Semantic embeddings genuinely beat keyword-only search on art archives — find ships by drawing style, not tags.
•90k → 1M image roadmap shows real data scope; discoverable public domain content solves a real friction point.
•No signup, open interface; clean gallery UX makes casual browsing feel good.

Weaknesses

•Public domain image search is a known category — Google Arts, Europeana, wikimedia Commons cover this.
•No evidence of retrieval accuracy vs. existing semantic search tools; 'aiming for 1M' is future work.

Post Description

With the increasing prevalence of AI-generated imagery, it's also hard to find what is real.

The public domain has many millions of interesting images, all free to use. Why not try to make them more discoverable. The site uses vector image embeddings to find content 'within' the image, rather than just in captions and titles. Search, for example, 'drawings of ships' and the semantically closest search results will be shown (along with lots of other features). 90k images currently added, aiming for 1 million soon.