Open-source project · Digital & AI
OpenVisualist
Context-aware image sourcing for the Public Domain. An open-source engine that reads a draft and sources real, public-domain photography to match it — no generative images, just provenance.
In an era of AI-hallucinated imagery, OpenVisualist prioritizes the provenance of the real. It reads an essay, understands what it's actually about, and sources genuine public-domain photography to match — for writers, historians and publishers who'd rather cite a source than fabricate one.
01 · The idea
Stop generating.
Start discovering.
Generative AI creates "synthetic" pixels. OpenVisualist acts as an autonomous librarian instead — reading the text, understanding the nuance, and sourcing real photography from history: Openverse, and eventually NASA and Wikimedia archives alongside it.
Every image carries a license shield — green for public domain (CC0/PDM, no attribution needed), amber for attribution-required (CC-BY) — plus a one-click "copy attribution" string, so nothing gets used without knowing exactly what it costs to use it.
02 · How it works
Write. Pause. Discover.
The interface is a split pane: a distraction-free Writing Zone on the left, a Curation Zone on the right. Three seconds after you stop typing, the engine reads the whole draft, expands what it finds into concrete visual search terms, and returns results grouped by the phrase that found them — each with a confidence meter and its license shield.
