Microcontrollers used in Project Axis smart glasses
Developing the Future

PROJECT AXIS

An initiative to design AI-powered smart glasses to help the 43 million people who are blind and the 2.2 billion more with vision impairment.

Explore Projects
SoundView smart glasses prototype for visually impaired users

Behind the Scenes

From the camera to audio feedback — our product is built with precision and carefully selected materials.

Read Blog
Project Axis upcoming project

Our History

From a school science project to an assistive technology startup — see how we got here.

Meet the Team
2.2B
People living with vision impairment worldwide
43M
People who are blind worldwide
4
Core components in SoundView v1
฿3,600
Total project budget, funded by Caleb's dad

[ Projects ]

Wind turbine vs solar panel energy comparison experiment

Wind Energy VS Sun Energy

Real-life engineering examples comparing wind turbine and solar panel energy output in a controlled test.

Read More
CO2 and air quality sensor monitor using microcontroller

CO₂ Levels

A smart environmental monitor tracking CO₂ levels and displaying real-time data using sensors and microcontrollers.

Read More
SoundView smart glasses prototype for blind users

Smart Glasses [Flagship]

AI-powered glasses for the 43 million people who are blind and the 2.2 billion more with vision impairment — minimal materials, affordable microcontroller hardware.

Read More
Project Axis blog posts with behind-the-scenes project updates

Blog Posts

Behind-the-scenes posts covering our projects — images, materials lists, and full reports.

Learn More
Animated preview of Project Axis hardware gallery

Gallery

A curated collection of images and captions documenting our builds and hardware.

Learn More
Project Axis team members Caleb, Binbin, and Raver

Information

Meet the team — their backgrounds, roles, and what drives the mission.

Learn More

How SoundView Works

A sound, a picture, a description.

01

Capture

A wide-angle camera on the frame continuously reads the space ahead — sidewalks, doorways, signage, faces, obstacles.

02

Process

The ESP32-S3 Sense records the spoken question via its onboard microphone and captures an image with its onboard camera, then sends both to OpenAI's Vision API over WiFi.

03

Describe

The scene is turned into a short, plain-language description — "curb ahead," "door on your left," "person approaching."

04

Feedback

The AI's text response is converted to audio and played back through a 13MM speaker mounted near the user's ear.

Who's Building It

A small, student-led team.

Project Budget

Project Axis runs on a ฿3,600 THB budget — funded entirely by Caleb's dad. Here's exactly how we're putting every baht to work building SoundView.

Use of Funds
Hardware Components (ESP32-S3, Speaker)67%
OpenAI API & Cloud Costs25%
Glasses Frame & Assembly5%
Jumper Wires & Accessories2%
Miscellaneous1%
Project Budget 2026
฿3,600
Total budget — Class 7A, 2026
Funded byCaleb's Dad
Hardware Components~฿2,400
OpenAI API Costs (est.)~฿900
Frame & Miscellaneous~฿300
Team Size3 Students

Questions

Frequently asked

What is Project Axis?

Project Axis is a student-led initiative building affordable, AI-powered smart glasses for people with vision impairment. We design, prototype, and document everything in the open as we go. It also includes many other projects, things that have to do with arduino, and IOT, so that we can learn from them and build the smartglasses.

How do the SoundView glasses work?

A camera on the frame reads the space ahead, OpenAI's Vision API identifies what matters in the scene, and the result is spoken back through a 13MM speaker driven by a Max98357A amplifier — see the "How SoundView Works" section above for the full breakdown.

Can I invest or support the project?

We welcome support of any kind — hardware donations, advice, or funding. Reach out at [email protected] or visit the Donors page to see how you can contribute.

Is the hardware or software open source?

Our builds are documented publicly on the Blog and Gallery pages, including materials lists and circuit notes, so others can learn from and build on our work.

Follow the build

One email whenever we ship something new — no spam, unsubscribe anytime.