AI navigation for blind people

WayFinder helps blind people move through the world with a phone camera, voice, and 3D audio

Find objects, detect obstacles, and navigate unfamiliar spaces through real-time scene understanding and spatial audio guidance.

Real-time obstacle alerts. Object finding by voice. Privacy-first processing.

A real-time navigation assistant built for the last few meters

Maps can get you to a street.
WayFinder helps with what happens next.

It turns a smartphone camera into a real-time navigation assistant for blind and low-vision users. Instead of describing the world in long paragraphs, it helps users move through it with spatial audio cues, voice alerts, and live object finding.

Moving safely still depends too much on guesswork

A white cane is essential, but it cannot always warn about a low sign, an open cabinet door, a scooter moving toward you, or where exactly the empty chair is.

For blind and low-vision people, this creates two constant problems:

  • Hidden obstacles that are hard to detect in time
  • Difficulty finding specific nearby objects in unfamiliar spaces

How WayFinder works

1

Point the phone forward

The user places the phone facing forward, for example on a chest mount, and starts the live camera view.

2

Ask for what you need

The user can say things like "Find an empty chair," "Where is the door?" or "Guide me around obstacles."

3

Follow the sound

WayFinder analyzes the scene in real time and responds with 3D audio cues, spoken confirmations, and urgent alerts for nearby risks.

See WayFinder in action

In the demo, a user asks WayFinder to find a free table, follows the audio guidance, and receives an immediate warning when a moving obstacle appears nearby.

Demo Video Coming Soon

Demo of object finding, obstacle alerts, and spatial audio guidance.

What WayFinder does best

Audio Compass

Continuous spatial audio guides the user toward the target without overloading them with constant speech.

Threat Detection

WayFinder detects nearby moving obstacles and warns the user before they become dangerous.

Find by Voice

Users can ask for nearby objects naturally, such as a chair, desk, door, or handle, and get guidance instantly.

Why this is different

Most navigation tools stop at maps.
Most AI vision tools stop at single images.

WayFinder is built for the missing layer in between: real-world movement at the last few meters.

Not just description. Real-time guidance.

It combines:

  • continuous scene understanding
  • live object finding
  • real-time safety alerts
  • spatial audio guidance

Privacy-first by design

WayFinder is built for real-world environments where privacy matters.

Video is processed in real time
Streams are not stored for training
Analysis is designed to minimize unnecessary data retention

Built as a working MVP

WayFinder already has a working mobile experience and a real-time demo pipeline for live video analysis, threat detection, object finding, and 3D audio guidance.

It is currently focused on improving latency, expanding real-world testing, and moving toward a more seamless wearable experience.

Live MVP

Help bring real-time AI navigation to the people who need it most

Watch the demo and join the beta waitlist.

Watch Demo