A traditional audio guide is a fixed script in a few languages, recorded once and expensive to change. Visitors arrive speaking dozens of languages and with questions the script does not answer: who was the sitter, why is it damaged, what else did she make. Meanwhile the curatorial knowledge to answer them sits in catalogues, files and labels the visitor never sees.
The guide answers from that knowledge. Curators supply the source material, such as labels, catalogue entries, exhibition text and approved research, and the guide retrieves from it to answer in the visitor’s language, by text or synthetic voice, through a phone browser or QR code with no app to install. If the sources do not answer a question, the guide says so and points to what it does know.
Interpretation is also an access service. The guide supports screen readers, adjustable text, plain-language and audio-description modes, built to WCAG 2.2 AA, so it serves visitors who are blind, partially sighted or have cognitive disabilities as well as visitors who speak another language.