The application works with open data provided by the Cleveland Museum of Art. After selecting a theme, several randomly chosen digital reproductions from the relevant collection are downloaded automatically.

The storytelling is based on factual data about the object, selected parameters about the visitor and the narrator, and system prompts that determine how to style and formally shape the story. Once the story has been processed by a language model, the resulting text is sent to a text-to-speech algorithm, which converts it into spoken word.

Cabinet Storyteller was created by Lukáš Pilka and the Cabinet of Wonders team in 2025.