The name is the problem. "Audio guide" tells deaf visitors that this product isn't for them. It's right there in the name — audio. Sound. The thing they don't have full access to.
But think about what's actually inside an audio guide: structured information about artworks, delivered stop by stop, in a sequence the museum designed. That content doesn't have to be sound. It can just as easily be text on a screen. The medium is incidental. The experience is what matters.
Most museums haven't thought about it this way. They treat accessibility as a separate workstream — produce the audio guide, then figure out how to accommodate everyone it doesn't serve. That approach is expensive, slow, and usually results in a watered-down alternative that nobody's happy with. A better approach: build the guide so the content works across modalities from the start.
Not one group, but three
Museums tend to lump everyone together. "Deaf and hard-of-hearing" becomes a single checkbox. In practice, these visitors have different needs, different preferences, and different relationships with technology.
Deaf from birth or early childhood. Many of these visitors use sign language as their primary language. Written text is a second language — usable but not always preferred. Visual content matters most. They're often comfortable with assistive technology and know exactly what they need.
Hard of hearing. This is the largest group, and the most varied. Some visitors can follow audio with the right volume and clarity. Others need text backup. Many are older adults experiencing age-related hearing loss who may not identify as "hard of hearing" at all. They just quietly stop using anything that requires good hearing.
Late-deafened adults. People who grew up hearing and lost their hearing later. They typically prefer text because their language foundation is auditory — they think in spoken language and read fluently. Sign language may not be part of their experience at all.
One accessibility solution won't serve all three groups. A sign language video overlay doesn't help the late-deafened visitor who reads English perfectly well. A text transcript doesn't fully serve the deaf visitor who'd prefer sign language. Flexibility is the only strategy that works.
The text-first audio guide
The most effective thing a museum can do is ensure that every piece of narration has a full text equivalent — not buried in a menu, not available on request, but right there as the default alternative.
This sounds obvious. It's not how most audio guides work. Traditional hardware devices are built around audio playback. You press a number, you listen. There's no screen, or if there is, it shows a title and maybe an image. The content lives in the audio track.
Phone-based guides, including modern AI audio guides, changed this, but many still treat text as secondary. The transcript exists somewhere in the app. You can find it. But the interface assumes you're listening.
A text-first approach means the guide works equally well with the sound off. The transcript isn't a fallback — it's a first-class way to experience the tour. The stop plays, and the visitor can listen, or read, or both. No separate workflow. No "accessibility mode" that strips out features. The same guide, used differently.
Musa built what we call karaoke mode for exactly this reason. As the audio narration plays, the full transcript appears on screen with real-time word-by-word highlighting — like reading lyrics while a song plays. Hard-of-hearing visitors can catch some of the audio while following along visually. Deaf visitors can ignore the audio entirely and still have an engaging, timed reading experience that preserves the pacing the museum intended.
It's a small feature that does a lot. The highlighting makes reading feel active rather than passive. You're not staring at a wall of text — you're following a guide.
Working with what visitors already carry
Something museums routinely overlook: the phones in your visitors' pockets already have sophisticated accessibility features. Both iOS and Android have spent years building hearing accessibility into their operating systems. Your audio guide doesn't need to reinvent this. It needs to work with it.
iOS offers Live Listen (which amplifies audio through AirPods or hearing aids), Made for iPhone hearing aid support, real-time captions for any audio, and LED flash alerts for notifications. Any well-built web or app-based guide that follows standard accessibility practices will plug into these features automatically.
Android provides Sound Amplifier, hearing aid streaming, live captions that work across all apps, and customizable notification settings including visual and vibration alerts. Same principle — build your guide correctly and Android does the heavy lifting.
That "build correctly" part is where most guides fall short. Many audio guide apps don't follow platform accessibility standards. They use custom audio players that bypass the system's captioning. They implement their own notification sounds that can't be redirected to visual alerts. They break the accessibility chain at the exact point where it matters.
When a guide does follow these standards, the layers stack up. A hard-of-hearing visitor using hearing aids connected to their iPhone gets amplified audio from the guide streamed directly to their ears, plus real-time captions on screen, plus visual notifications when it's time to move to the next stop. All of that works because the guide plays nicely with the platform. We've seen this combination in practice — visitors get a richer experience than either the guide or the device could provide alone.
Typed interaction instead of spoken
Most AI-powered guides let visitors ask questions by voice. Tap the microphone, speak your question, get an answer. This is great for hearing visitors. For deaf visitors, it's another barrier.
The fix is straightforward: let people type. A text input field that accepts questions the same way the microphone does. Same knowledge base, same answers, same experience — just a different input method.
This matters more than it seems. The ability to ask questions is what separates a modern AI guide from a static audio tour. If that feature is gated behind speech input, you've built an interactive guide that's only interactive for people who can speak and hear. Typed input keeps the full experience available to everyone.
There's a secondary benefit too. Some hearing visitors prefer typing in quiet galleries. Parents with sleeping children in strollers. People who are self-conscious about talking aloud in a museum. Text input serves accessibility and preference at the same time.
Sign language: the honest picture
Sign language video integration is the gold standard for deaf visitors who use sign language as their primary language. A signer appearing on screen at each stop, delivering the interpretation in BSL, ASL, or the local sign language — that's the most accessible option for this group.
It's also expensive and logistically difficult. Sign language isn't universal. BSL and ASL are completely different languages. A museum serving international visitors would need multiple sign languages, just as it needs multiple spoken languages. Each requires a qualified interpreter, video production for every stop, and updates whenever the exhibition changes.
Some museums do this well. They invest in sign language videos for permanent collections and accept the cost as part of their accessibility commitment. That's worth doing.
For most museums, though, sign language video for every stop in every relevant sign language isn't realistic right now. What is realistic: offering sign language for the most popular stops or permanent highlights, while ensuring the text-based experience is strong enough to serve deaf visitors on everything else. Prioritize rather than pretend you can cover everything.
AI-generated sign language is on the horizon. Avatar-based signing has improved significantly, though it's not yet natural enough for most deaf communities to find it comfortable. This will change, and when it does, it'll solve the scaling problem — just as AI solved the spoken language scaling problem for audio. But it's not ready today, and overselling it would be dishonest.
Visual notifications and navigation
Sound carries information beyond narration. A chime that tells you to move to the next stop. An alert that the museum is closing in fifteen minutes. Background audio cues that shift as you enter different galleries.
Deaf visitors miss all of this unless you've designed visual equivalents.
Haptic feedback. A brief vibration when it's time to move on. Most phones support this and it's trivial to implement. Yet many guide apps don't.
On-screen indicators. A visual pulse or banner that replaces audio cues. "Next stop is to your left" as text, not just as spoken instruction.
Progress indicators. Where am I in the tour? How many stops remain? Hearing visitors pick this up from conversational cues in the narration ("we're now halfway through our visit"). Deaf visitors reading text deserve the same orientation.
These aren't expensive features. They're design decisions that someone has to remember to make. The problem isn't technical difficulty — it's that guide designers can hear, and they test the product with the sound on.
What museums should do
If you're evaluating your museum's audio guide accessibility, here's a practical checklist. Not everything is equally urgent, but the first three are non-negotiable.
- Always offer a full text transcript for every narration. Not buried in settings. Right there on the main screen. If a visitor turns off their phone's volume, the guide should still be fully functional.
- Support typed input for questions. If your guide has any interactive or AI-powered features, text input must exist alongside voice input. Otherwise you've built interactivity that excludes the people who need it most.
- Follow platform accessibility standards. Use native audio players, standard notification patterns, and system font sizing. Let iOS and Android do what they're built to do. Don't override system accessibility settings with custom implementations.
- Add visual and haptic notifications. Every audio-only alert should have a visual or vibration equivalent. Test the entire guide with sound muted.
- Provide captions for any video content. If your guide includes video clips — artist interviews, conservation footage, historical material — caption them. Auto-generated captions are a starting point, not a finished product. Have someone review them.
- Consider sign language for key stops. You probably can't cover everything, but covering the top ten stops in the local sign language sends a strong signal and serves the visitors who need it most.
- Ask deaf visitors what works. Partner with local deaf organizations during development, not after launch. Their feedback will catch problems you can't see because you can hear.
The name is still the problem
We've been building better text support, better device integration, and better interaction options into audio guides — but we haven't changed what we call them. "Audio guide" still signals to deaf visitors that this isn't for them. Some museums have started using "multimedia guide" or just "museum guide." That's a small change with real signaling value.
The content inside these systems has moved beyond audio. It's text, images, maps, interaction, and yes, sound — but sound is one component, not the whole product. The name should reflect that, especially when you're trying to tell visitors with hearing loss that this experience was built for them too.
If your museum is working on making its guide accessible to deaf and hard-of-hearing visitors, we can share what we've learned. This is an area where small design decisions make a large difference, and getting the details right matters.