Seba and Smartify

Seba lets a visitor identify an exhibit by photographing it — no QR code, no app download, no account creation. It then plays a short spoken introduction and lets the visitor ask follow-up questions in a conversation.

PUAM already has a digital guide in Smartify, launched with the Museum’s reopening. This isn’t a pitch to replace it — it’s a look at what a photo-based, conversational layer could add, particularly for the densely displayed cases.

Which Smartify? Smartify comes in two forms:

An installable app that visitors download and set up on their mobile device. Most casual visitors won’t do this in the middle of a visit; it works best where the museum lends tablets and headphones.
A web app that opens in the mobile device’s browser — PUAM’s is at princetonart.smartify.org.

Seba is also a web app, so this comparison is with PUAM’s Smartify web app.

How they compare

	Seba	Smartify web app at PUAM
Exhibit recognition	Recognizes the object itself from a visitor’s photo, matched against reference photos taken by staff — no marker needed	QR codes placed in the galleries, currently on a limited number of objects. (Smartify’s image recognition is available only in the installable app.)
Visitor interaction	Spoken intro, then open-ended follow-up conversation	Label text and provenance, on request
Personalization	Age group, knowledge level, length, language	Not evident in the current deployment
Platform cost	No license fee; small per-use AI costs. Can run on PUAM’s own servers	Recurring annual subscription fees
Visitor data	No accounts or personal details. An anonymous cookie supports usage statistics (exhibits viewed, time spent), kept on the Museum’s server	Accounts, cookies, and engagement messaging referenced in Smartify’s own materials
Where data lives	Content and usage statistics on Princeton’s servers; visitor photos and questions are processed by AI providers (Anthropic, Voyage AI)	Smartify’s infrastructure
Feature iteration speed	Quick changes — the code can be Princeton’s own	Subject to vendor roadmap and support cycle
Adding and editing exhibits	PUAM staff add an exhibit or edit its description in minutes, from a mobile device or computer	Through Smartify’s content management system

Where Seba adds something new

No QR code on the exhibits. Densely packed cases may not have room for QR codes. Seba recognizes the object itself, so there are no codes to print, place, or replace when a gallery is rehung.
A visitor can actually ask questions. Today’s guide delivers label text and provenance. Seba has a real back-and-forth — “why does this matter,” “what else is like this” — with answers grounded in the Museum’s own material.
Staff keep it current in minutes. Museum staff add an exhibit by photographing it, and write or edit its description, from a mobile device or computer — no vendor involved. Built-in analytics show which exhibits visitors view and for how long, and point to exhibits that are hard to recognize so staff can add more reference photos.
It adjusts to who’s standing there. Age group and knowledge level, set by the visitor. A scholar and a ten-year-old get different answers to the same exhibit.
The Museum stays in control. No visitor accounts and no personal information. Usage statistics stay on the Museum’s server; the only outside services are the AI providers that process photos and questions. The Museum owns the code, the content, and the visitor’s experience of both.

What Smartify does well

Reach. Over 1,200 partner organizations and more than six million users — a visitor may already have the app from another museum, though that advantage applies mainly to the installable app.
Breadth. Maps, wayfinding, accessibility features (audio description, sign-language tours), and e-commerce — none of which Seba attempts today.
It’s already live. Staff are trained on it and content is already in place. Seba is a proof of concept, not yet a production system.

Smartify figures from smartify.org, October 2026.