AR has been chasing the wrong sensor.

The world is not just visible. It’s also audible and radio-bright. Every room you walk into is full of structured signals: the way sound bounces off walls and furniture, the way Wi-Fi packets bend and attenuate as they pass through people. These signals describe the space with extraordinary precision — and unlike cameras, they work through walls, in the dark, and without recording the people in front of them.

The science has been quietly catching up. Acoustic SLAM — using inaudible chirps from a phone’s own speaker and microphone — has hit 0.1m to 0.5m median accuracy on commodity hardware in peer-reviewed work. Wi-Fi Channel State Information has been used to detect motion, presence, and breathing through drywall. Last year, the IEEE published 802.11bf: the first standard to bake sensing primitives directly into the Wi-Fi spec. Routers shipping in the next two years will carry it.

The platform layer is unowned.

A new sensing modality always works the same way. First, the physics is demonstrated in research. Then, slowly, hardware catches up. Eventually, somebody builds the software layer that makes it usable to a working developer — and that somebody usually defines the category for a decade.

For LiDAR-based spatial computing, that layer was ARKit and ARCore. For GPS, it was the carrier APIs and then the operating system. For Bluetooth ranging, it was Apple’s Find My network. None of those layers exist yet for non-optical spatial sensing. There is no equivalent of CoreLocation for “tell me where I am, indoors, to within half a metre, using only the speaker and microphone in this phone.” There is no WiFiSensingKit.

That’s the gap acoAR fills.

What we’re building.

acoAR is a cross-platform SDK for spatial sensing without cameras or LiDAR. It runs on iOS and Android, talks to the speaker, microphone, IMU, and (where available) Wi-Fi radio, and gives developers three primitives:

• Position. Sub-meter indoor location, with no beacons and no fingerprint database to maintain.
• Room. Identify which room a device is in by its acoustic signature — like Shazam for spaces.
• Presence. As 802.11bf chipsets ship, detect occupancy and motion through walls, with the user’s consent and without storing anything that resembles a person.

Three modalities, fused into a single spatial stream that any application can subscribe to. No new hardware to install. No cameras pointed at people who didn’t agree to be seen.

Why this matters now.

There is a class of problem — eldercare in private bathrooms, retail traffic in changing rooms, security in schools, robotics in low-light spaces — where a camera is either illegal, impolite, or simply doesn’t work. These markets have spent the last decade waiting for a sensing modality that respects the constraint. They’ll spend the next decade buying it.

And there is a class of developer — mobile, AR, robotics — who has been told that real spatial awareness requires a flagship phone with a depth sensor or a $1,500 headset. We think that’s a temporary state of affairs. We think the phone in your pocket already has the sensors, and what’s been missing is the software.

What we’re not.

We are not an AR glasses company. We are not selling a hardware puck. We are not replacing LiDAR — LiDAR is excellent at what it does, and a phone with LiDAR can use both. We are building the layer underneath: the part that turns sound and radio into a coordinate stream a developer can read.

We don’t want to be the consumer brand. We want to be the infrastructure that other consumer brands quietly run on.

What you can do.

If you build mobile, AR, or robotics software: join the waitlist for SDK access in Q3 2026. We’ll send one email when it opens. No drip campaign.

If you’re working on a deployment in eldercare, smart buildings, retail, or wayfinding: talk to us. We’re partnering with a small number of teams ahead of public launch.

And if you just think this is the right shape of bet: hello@acoar.com.

— the acoAR team, London