A Roman villa site in the countryside. A medieval castle ruin on a hilltop. A prehistoric stone circle in a field with no buildings for a mile. These places have stories worth telling, and nobody to tell them.
The challenge is practical, not intellectual. There's no Wi-Fi. Often no cell signal. Sometimes no power supply. The site might not even have a permanent structure to mount a sign on. Traditional audio guide solutions assume a controlled indoor environment with reliable connectivity. Heritage sites in the wild don't offer that.
So most of them go without. Maybe a few interpretation panels that visitors half-read. Maybe a printed leaflet from the gift shop. The actual experience of standing where history happened, with context that makes it meaningful — that's left to the visitors who happened to read the right book beforehand.
It doesn't have to be this way.
The infrastructure assumption
The audio guide industry was built for museums. Climate-controlled buildings with power outlets, Wi-Fi routers, and a front desk to hand out devices. The technology developed around those assumptions.
Beacon-based systems are the clearest example. You install Bluetooth beacons near each exhibit, and when a visitor's phone detects one, it triggers the right content. This works well indoors. For an outdoor heritage site covering ten acres of open ground, it's absurd. You'd need dozens of beacons, each requiring batteries or power, each exposed to weather, each needing maintenance. Some vendors quote this with a straight face. The cost is high, the maintenance is ongoing, and the first harsh winter takes out half your network.
QR codes at every stop are the low-tech alternative. Print a code, mount it on a post, let visitors scan as they walk. This works in theory. In practice, it means visitors are constantly pulling out their phone, opening the camera, scanning, waiting for the page to load — assuming they have signal. By the third stop, the novelty has worn off. By the fifth, they've stopped scanning. It turns interpretation into a chore.
Both approaches also assume someone is maintaining physical installations across the site. For a castle ruin managed by a volunteer trust with an annual budget in the low thousands, that's not realistic.
BYOD is the only option that scales
For outdoor heritage sites, bring-your-own-device isn't just a nice option. It's the only approach that makes sense, and it's the foundation of how a modern AI museum guide works.
No hardware to buy, install, maintain, or replace. No beacons to weatherproof. No devices to charge, sanitize, and track. The visitor's phone is the guide. The site's only physical requirement is a QR code at the entrance — one code, not one per stop — and maybe a sign explaining that a guide is available.
The economics are straightforward. A site that can't justify a five-figure hardware investment can still offer a guided experience. The ongoing cost is the software and content, not replacing stolen handsets or swapping beacon batteries every six months.
But BYOD for outdoor heritage sites hits a wall that indoor museums rarely face: what happens when the visitor's phone can't reach the internet?
Making offline actually work
An offline audio guide isn't just a regular guide with the Wi-Fi turned off. The whole interaction model has to be designed for it.
With Musa, a visitor arrives at the site and scans a single QR code. That triggers a content download — the tour data, the narration, the floor plans or maps, the images. Once that's on their phone, they don't need to scan anything else. Ever. The entire tour runs locally.
That's a big difference from the "QR code at every stop" model. One scan, then the phone guides you through twenty stops without touching the camera again. The curated tour path works. Inline suggestions work — the guide can still say "the foundation stones behind you are from the original structure, tap here to learn more." Diverging from the tour path works. It's the same guided experience, just running on cached content instead of live data.
What doesn't work offline: asking the AI custom questions. If a visitor wants to say "what was daily life like for the servants who lived here?" that requires a round trip to the AI model, which requires a connection. The curated content and pre-built tour logic handles a lot — but the open-ended conversational layer needs internet.
We're honest about this limitation. It's a real tradeoff. But consider what visitors actually do on a curated tour. Most of them follow the path, listen to the narration, maybe tap a suggestion or two. The percentage who fire up a custom question at an outdoor ruin with wind in their ears is small. The offline experience covers the vast majority of what people want.
How visitors find what they're looking at
Indoor museums have it easy. You can label every object, number every room, put a map on the wall. Outdoor heritage sites are messier. A visitor standing in a field of ruins might not know which pile of stones is the kitchen and which is the chapel.
Navigation in these environments uses two complementary approaches.
GPS positioning handles the broad strokes. Outdoors, phone GPS is accurate enough to know which part of the site a visitor is standing in. The guide can track their position and trigger content for the nearest point of interest. No beacons, no scanning — the visitor walks and the guide follows. GPS doesn't work indoors (the signal degrades inside buildings), but for open-air heritage sites, it's free infrastructure that every smartphone already has.
Visual anchor points handle the fine detail. These are reference photos — images of specific landmarks, architectural features, or trail markers that help visitors confirm they're in the right spot. "You should see a stone archway ahead" is more useful than a GPS coordinate. For indoor sections of a site (a crypt, a chapter house, a surviving building), visual anchors plus floor plans replace GPS entirely. Visitors match what they see on their screen to what they see in front of them.
Neither approach requires installing anything at the site. GPS is satellite-based. Visual anchors are content, not hardware — photos and descriptions that live in the app alongside the tour narration. A site manager adds them once during setup and they work forever.
This is the same system Musa uses for indoor location tracking at museum partner sites — visual anchor points and floor plans, no beacons. The outdoor layer adds GPS on top. The technology is the same; the environment just changes which positioning method takes the lead.
What about sites with partial connectivity?
Not every heritage site is a total dead zone. Many have patchy coverage — maybe there's signal in the car park and the visitor center, but it drops out on the trail to the far ruins. Some sites have Wi-Fi in the main building but nowhere else.
This is actually fine. The download happens wherever there's a connection — the car park, the visitor center, the ticket office. Once the content is on the phone, the visitor can walk into complete signal blackout and the tour keeps running. If they wander back into a coverage area, custom questions become available again.
The practical advice: put your QR code where the signal is. If the visitor center has Wi-Fi, put the code there. If the car park is the only spot with cell coverage, put the code in the car park. The download takes less than a minute for a typical tour. After that, the visitor is untethered.
One QR code, not thirty
This is the single biggest UX difference from the "QR code at every stop" model that most heritage sites default to.
Scanning a QR code is a small friction. Doing it once is fine. Doing it at every stop turns your heritage experience into a scavenger hunt for little black-and-white squares. The visitor is looking at their phone instead of looking at a 900-year-old wall. Each scan is a moment where they might decide it's not worth the effort. And if one code is damaged, faded, or missing, that stop has no interpretation at all.
One scan at the start. Then the guide takes over. If the visitor wants to jump to a specific point of interest out of order, they type its name. That's faster than walking to the right QR code, more reliable, and works even when the visitor can't see a physical marker. It's a small design choice with a big impact on whether people actually use the guide for the whole visit.
Honest state of outdoor heritage guides
We should be upfront about where things stand. Musa currently works with indoor partner sites — museums, galleries, historic houses. We don't yet have a fully outdoor heritage site live on the platform. One partner is actively planning an outdoor tour, and the technology (GPS, visual anchors, offline mode) supports it. But it hasn't been battle-tested across dozens of outdoor sites the way our indoor product has.
That said, the building blocks are proven. GPS is a known quantity. Offline content delivery is a solved problem. Visual anchor points work the same way whether the anchor is a painting in a gallery or a stone gatehouse in a field. The gap isn't technical capability — it's field experience with the specific challenges of outdoor sites. Wind noise, rain, sun glare on screens, visitors in heavy gloves who can't tap easily. These are real problems that get solved by iteration, not by pretending they don't exist.
If you manage an outdoor heritage site and you're evaluating audio guide options, the honest picture is: BYOD with offline content and GPS navigation works today. The AI conversational layer works wherever there's connectivity and degrades gracefully where there isn't. Full offline custom Q&A is the last missing piece, limited by the fact that running a large language model on a phone isn't practical yet. What you get offline — curated tours, suggestions, navigation — covers the core experience.
For site managers
If you're responsible for interpretation at an outdoor heritage site, infrastructure shouldn't be your blocker anymore.
You don't need to wire the site for Wi-Fi. You don't need to budget for beacons. You don't need to install and maintain QR codes at every point of interest. You need content (the stories, the history, the interpretive narrative), a single entry point for visitors to access it, and a system that delivers it without depending on connectivity after that first moment.
The content is the hard part, and it always has been. Not the technology. A Roman fort with a deeply researched, well-narrated tour will outperform a museum with a boring one, regardless of how much hardware either site has. The technology just needs to get out of the way and let the content reach the visitor.
If you're thinking about audio guide options for a site without traditional infrastructure, get in touch and we can talk through what's realistic for your specific situation.