Technology & Implementation7 min read
If you're running a museum or heritage site, you've probably heard of Nubart. Their physical card system has a nice appeal: visitors buy a card with a unique QR code, scan it on their phone, and access a web-based audio guide. No app to download, no friction. But physical cards come with real tradeoffs that don't work for every venue.
Nubart's model was clever for 2018. Today, the constraints of physical inventory, limited per-visitor data, and static content feel increasingly dated. If you're evaluating audio guide platforms, you should understand what you're actually trading away when you go the card route—and what's available now that wasn't an option five years ago.
Nubart prints physical cards, each with a unique QR code printed on it. Visitors buy the card (or grab it at your desk) and scan the code to access a web-based guide. The card itself becomes a souvenir. From a curator's perspective, you upload content to their platform, and the system serves it to whoever scans that card's code.
The appeal is obvious: no app install friction, no membership, no registration unless the visitor wants to. The card acts as a commitment device—people are more likely to engage with something they've paid for or picked up physically.
The limitations are more subtle. Because each card is static, you're limited in what you can track. You see how many people scanned the card, but not much about their behavior. If you want to understand which stops people actually listened to, how long they spent, whether they skipped content—that data is sparse or unavailable. You also can't customize the experience per visitor. If someone has mobility concerns or speaks a language Nubart hasn't added yet, tough luck.
And physical inventory is inventory. You need to manage stock, reorder cards, handle damage and loss, and deal with the logistics of distribution.
Inventory management. Cards get lost, damaged, or stolen. If you're scaling to multiple languages or creating cards for traveling exhibitions, you're managing stock levels across different versions. That's overhead.
Limited behavioral data. Nubart gives you scan counts and maybe basic engagement metrics, but not per-visitor journey data. You don't know which stops people actually consumed, how long they listened, or whether they came back. This makes it hard to iterate on content based on real visitor behavior.
Static content and no personalization. Once a card is printed, its QR code points to a fixed version of your guide. You can update the web content, sure, but you can't serve different experiences to different visitors—no A/B testing, no dynamic tours, no adaptive difficulty for families with kids.
No conversational layer. A card-based system is fundamentally passive. Visitors scan, they listen, they move on. There's no back-and-forth, no ability to ask a question or get clarification.
Language and accessibility limits. You're locked into whatever Nubart supports. If you want to add a new language or need advanced accessibility features, you're waiting for their roadmap.
The audio guide landscape has shifted. Smartphones are ubiquitous and expected to handle everything. Visitors don't balk at pulling out their phone—they're already doing it for maps and photos. QR codes, after years of niche status, are now normal.
Most importantly, the AI tooling that makes interactive, conversational guides possible is now accessible to operators at any scale. You don't need to be a tech company to offer an experience that adapts to how visitors actually behave.
The simplest alternative is to abandon physical cards and go QR-only. Print your code on signage, include it in visitor materials, or email it. You get all the benefits of Nubart's frictionless model without inventory overhead.
Platforms like Musa let you do this: a single QR code, visitors scan on arrival, no app required, web-based guide. You control the code, so you can update where it points without reprinting anything. You get full per-visitor analytics—which stops were visited, dwell time, return visitors, language preferences. You can update content on the fly.
The tradeoff: visitors don't walk away with a physical souvenir. This matters to some venues and not at all to others. For venues focused on conversational guides or where multiple visits are expected (members, locals, repeat tourists), the analytics value usually outweighs the loss of the card.
The real shift is toward AI audio guide platforms that treat audio guides as interactive experiences, not content containers. These platforms typically include:
Musa is built this way. Curators create a knowledge graph—objects, artists, themes, connections. Visitors scan a QR code on arrival and can navigate however they want. They can ask questions and get answers grounded in what you've taught the AI. The system learns which content resonates and flags what might need updating.
There's no physical inventory. There's no per-venue QR code limit. Multiple museums can use the same platform and their data stays separate.
If you want full control over your stack and visitor data, self-hosted or open-source audio guide platforms exist. These typically require technical implementation (hosting, database, API setup) but give you complete flexibility. The tradeoff is operational overhead—you're responsible for keeping the system running, backing up data, and adding features.
These work well for large institutions with in-house tech teams. For most venues, the operational burden isn't worth the control gain.
This isn't a hit piece on Nubart. Their system works well for specific use cases:
For heritage sites with a single main path and clear monetization around merchandise, Nubart can make sense. For venues that want to understand visitor behavior, offer multiple languages, or let people explore dynamically, it's a constraint.
When you're evaluating vendors, think beyond the upfront cost:
Cost per visitor. Nubart's cards have material and printing costs. QR-only systems typically charge per visitor or per month. Do the math for your volume. A 200,000-visitor-per-year museum might pay the same monthly fee as a 20,000-visitor venue on Musa (usage-based pricing), but the per-visitor cost is vastly different.
Operational overhead. Include staff time managing inventory, reprinting, updating content. A platform that doesn't require new physical assets printed is significantly cheaper to operate once you account for labor.
Data ownership. Can you export your visitor data? If you leave the platform, can you take your analytics with you? This matters more than it sounds when you're making strategic decisions about your space.
Speed to new languages. How fast can you reach Spanish-speaking, Mandarin-speaking, or Arabic-speaking visitors? A platform with instant translation is months ahead of vendors who rely on manual localization.
Integration with your ops. Do they handle payments? Timed access (peak/off-peak rates)? Memberships? Analytics dashboards you can actually use, or just a CSV dump?
When you're evaluating alternatives to Nubart:
Most venues don't need to revolutionize their visitor experience overnight. If your Nubart setup is working and you're not feeling the pain of inventory management or limited data, stick with it. Migration friction is real.
But if you're setting up a new venue, expanding to a new space, or adding additional languages, you're in an ideal position to evaluate modern alternatives. The QR-code-plus-conversational-AI model isn't theoretical anymore—it's cheaper to operate, gives you better data, and genuinely works for visitors.
The card era was real. It solved a problem. But the problem it solved is mostly solved now by simpler means, and the problems it created (inventory, static content, limited analytics) are worth reconsidering.
Do I really need conversational AI for an audio guide?
Not necessarily. Conversational AI is valuable if your audience wants to explore non-linearly or ask follow-up questions. For highly structured venues or shorter tours, a linear guide is fine. But if you're operating a large museum where visitors have different interests and skip around, conversation dramatically improves the experience. Most visitors prefer asking a question to finding the right menu option.
Can I print QR codes on my existing materials instead of selling cards?
Absolutely. If you're not monetizing the guide itself (only your core admission or membership), printing a static QR code on signage, in brochures, or via email is the move. One code, no inventory, same frictionless experience as Nubart but with zero overhead.
How much does real visitor analytics actually change what you do?
More than you'd think. When you see that 80% of visitors skip a particular stop, or that three specific artworks drive questions, you optimize. You spend curatorial effort where it matters. Without data, you're guessing. With it, you're iterating.
What about offline guides? Don't visitors need internet?
For audio guides, internet dependency is now a feature, not a bug. Visitors expect to use their phones anyway, and downloading hundreds of MB of audio for offline use is clunky. Most venues are on reasonably solid WiFi. If offline is a hard requirement, you're in a narrow use case—look for platforms with offline download options, but know that limits your ability to update content.
readTimeMinutes: 7 audience: b2b
If you're building a modern visitor experience and looking for a platform that goes beyond card-based limitations, contact us to see how Musa handles guided discovery at scale.