Demo

Enter anything

Get a collage

It scrapes images, text, and archival material from Openverse, Wikipedia, Hacker News, Reddit, the Wayback Machine, Wikimedia, Wikiquote, the Internet Archive, MusicBrainz, and patents, and assembles them into a collage.

Single words work, but can be ambiguous

Ambiguous queries drift across domains

Specificity unlocks depth

Architecture

Frontend submits a job -> Celery workers scrape topic-related content via FastAPI and Redis -> Frontend composes the collage.