LinkedCulture now narrates its collection.

LinkedCulture now narrates its collection.

The entire process is automated. LinkedCulture pulls the object image and institutional metadata, Claude writes a four-sentence description, naming the technique, describing what is visible, grounding it in provenance, and explaining why the object exists. That description becomes the closed captions AND the narration. Kokoro TTS converts it to speech, FFMPEG builds the 4K video, and it posts automatically to YouTube.

Institutionally curated metadata combined with vision ML produces descriptions grounded in both the object's history and what is actually visible. The result is more complete than either source alone.

The video attached is one example from the live collection. 25 are published so far:

A Rembrandt etching of the Dutch countryside, c. 1650. A Cuzco School painting in oil and gold leaf, colonial Peru, ca. 1770. Roger Fenton's salted paper prints from the Crimean War, 1855. Eadweard Muybridge at Yosemite, 1868. Edo-period woodblocks, George Bellows lithographs, 17th-century etchings.

Built for accessibility and discovery. An embeddable player version is in the works for institutions who want to surface these directly in their own interfaces.

Full playlist on YouTube: https://lnkd.in/eUNDVEdg

Rolling out video on a regular basis...is this when I am supposed to say, like and subscribe?

#LinkedCulture #OpenAccess #CulturalHeritage #MuseumTech #Accessibility

The Metropolitan Museum of Art

Anthropic