Open Access
Published:
In 2017 the Met launched Open Access: images of every public-domain work in the collection, and data for every object, released under Creative Commons Zero for anyone to use.
The collection data lived in proprietary systems built to hold records, not serve the public. We wanted that data out quickly both to see how people would use it and to inform subsequent iterations.
The first release was a CSV of the whole collection, published on GitHub alongside the announcement. It was the quickest win, a mostly flat export that let researchers analyze the collection in bulk.
In 2018 my team launched the Collection API so partners, students, independent developers, and art tech nerds could query the collection directly.
In 2020 the API was serving over 3 million requests a month. By mid-2026 it had served over 900 million.
Who used it
- Wikipedia and Wikidata - Met artworks drew about 225 million views a year on Wikipedia by 2020.
- Met x Microsoft x MIT - AI prototypes built on the Open Access data.
- Parsons data visualization - students working directly with the API.
Further reading: Celebrating Three Years of Open Access at The Met (2020, with Sofie Andersen) and Connecting to art (Increment, 2020).