compareeHub

TURN ANY GLASSES INTO AI SMART GLASSES FOR UNDER 25 DOLLARS

Under 25 dollars of parts — a tiny ESP32 camera board clipped onto any glasses — and an AI model that answers questions about what you saw.

by BasedHardware

FULL CAD BOM FIRMWARE DOCS

AIOpen-hardware

Built withESP32Arduino3D printing

difficulty
●●●○○
time
a weekend
license
MIT
repo
repo UNMAINTAINED (MOVED TO OMI)4,175 stars

WHAT YOU’LL NEED

Videos
1
Jump to section

COMPAREE VERDICT

OpenGlass came out of a Meta Llama 3 hackathon and the idea is exactly what it sounds like: a XIAO ESP32-S3 Sense (camera plus mic, about 15 dollars) in a 3D-printed clip that sits on your glasses, takes photos, and lets an AI model (Llama 3 via Groq, OpenAI, or a local Ollama model) answer questions about what you saw. The whole build is under 25 dollars if you already own a 3D printer. The firmware is MIT-licensed and flashes via Arduino IDE. The hard part is not the hardware — it is the software side: you run an Expo app on your computer (currently as a web app), add Groq and OpenAI keys, and pair it with the glasses. The README lists the steps and assumes you know what 'clone the repo and npm start' means. Note that BasedHardware has moved development to its Omi project and marks this repo as no longer supported, so expect to fix outdated dependencies yourself. The camera is a tiny fixed-focus module, so image quality is closer to a webcam than a phone. The specified battery is a small 250 mAh LiPo, so runtime is limited. If you are comfortable flashing ESP32 boards and running a Node project, this is a genuinely clever weekend build. If you have never done either, you will spend most of that weekend troubleshooting the software stack, not the glasses.

GOOD TO KNOW

  • —MIT license — commercial use is fine.
  • —The 3D file for the mount is linked from the README (hosted outside the repo); firmware flashes via Arduino IDE.
  • —You need Groq and OpenAI API keys (or a self-hosted Ollama server) and the companion app, which runs as an Expo web app on your computer.
  • —The repo contains the firmware and an Expo app you run yourself; the only sign-ups are the Groq/OpenAI API keys.
  • —The camera is fixed-focus and low-resolution; this is not Google Glass-grade optics.
  • —No assembly guide — the README assumes you know how to solder headers and flash an ESP32.

Parts to buy

4 items

From our check of the build. Exact quantities and part numbers are in the creator’s BOM.

  • Seeed Studio XIAO ESP32-S3 Sense (about 15 dollars)Find
  • EEMB LP502030 3.7 V 250 mAh LiPo batteryFind
  • Optional: tactile buttonFind
  • WiresFind

BUILDS OF THE WEEK

Five open-source builds worth your weekend, every week.

Checked like this one: what’s really in the repo, what it costs, how hard it is. One email, unsubscribe anytime.

Can I build this?

PrintOne clip mount for your glasses frame (STL linked from the README, prints in 1-2 hours)
BuySeeed Studio XIAO ESP32-S3 Sense (about 15 dollars), EEMB LP502030 3.7 V 250 mAh LiPo battery, optional: tactile button, wires
Tools3D printer, soldering iron, Arduino IDE, a computer with Node.js to run the Expo app, Groq and OpenAI API keys (or a local Ollama server)
SkillsIntermediate: soldering headers, flashing ESP32 firmware, running a local backend server, basic terminal commands. Not a first electronics project.
TimeA weekend — half a day for the hardware, the rest for getting the server and app talking to each other
CostUnder 25 dollars in off-the-shelf parts according to the README, dominated by the XIAO ESP32-S3 Sense board; the LiPo battery and the printed mount are the rest.
SafetyLithium battery — do not short the leads, do not puncture, keep away from children. Standard soldering iron care.

Build at your own risk. Projects involve tools, electronics and sometimes mains voltage — follow the creator’s safety notes.

Videos

OpenGlass Demo - Meta AI LLaMa Hackathon - 12th May 2024

The official demo from the Meta Llama 3 hackathon, linked from the README. It shows the glasses in use rather than a step-by-step build.

More builds like this

All projects
A 28.9M parameter language model runs entirely on a cheap ESP32-S3 microcontrollerAIBuild a talking AI assistant on a breadboard and host its brain yourselfAIYou can build the motion capture rig big-budget films useOpen-hardware

Gallery

opengraph.githubassets.com
radar.comparee.ai
img.youtube.com

Start here

Navigation into the creator’s own docs — we don’t rewrite the guide, we route you to the source.

  1. 1.Read the hardware setup section in the README (Lists the exact XIAO board model and battery specs)
  2. 2.Flash the firmware via Arduino IDE (Install the ESP32 board package first, then upload the sketch)
  3. 3.Install and start the Expo companion app on your computer (Requires Node.js, npm or yarn, and Groq/OpenAI API keys (or a local Ollama server))

Resources

Documentation, files and community threads for this build — we link straight to the original sources and never rehost the creator’s files.

KNOWN ISSUES

  • The XIAO ESP32-S3 Sense is a specific board variant — the plain ESP32-S3 does not have the camera module. Ordering the wrong one wastes a week and 15 dollars.
  • The battery has to be soldered to the XIAO's battery pads on the underside of the board; get the polarity right and insulate the joints, because a short on a LiPo is dangerous.
  • The companion app is an Expo project that the README runs as a local web page; keep the computer within Bluetooth range of the glasses, and put your Groq/OpenAI keys or Ollama URL into keys.ts before starting it.
  • The Groq and OpenAI APIs have rate limits and usage costs — heavy use (constant photo capture and queries) adds up quickly. Self-hosting Ollama avoids API costs.
  • The camera auto-captures every few seconds by default, which drains the battery fast. You will want to tweak the interval in the firmware for real-world use.
  • There is a single STL for the mount and no sizing notes; test-fit it on your frames and modify the model if it does not clip on securely.

Do I need to keep my phone nearby?

Not your phone, but a computer: the glasses send photos and audio over Bluetooth to the companion app, which the README runs as a local web page on your computer, so keep that computer within Bluetooth range. The app calls Groq, OpenAI or your own Ollama server; the ESP32 does not run the AI itself.

Can I use a different AI model instead of Llama 3?

Yes — the app already has modules for Groq (Llama 3), OpenAI and a self-hosted Ollama server, so you can switch to a local model to avoid API costs.

How good is the face recognition?

The repo does not document a face-recognition feature or any accuracy figures; 'remember people you meet' is a stated goal. The camera is a small fixed-focus module, and recognition quality depends on the AI model you connect, so test it yourself.

Is there a way to turn off the camera but keep the mic?

The firmware already lets the app stop photo capture or take single photos over Bluetooth; the stock app starts periodic capture, so you would change the app (or the firmware) to keep only the microphone.

Community builds

No community builds yet — be the first, we feature the best ones.

Discussion1

FROM THE COMPAREE TEAM

The whole thing runs on a tiny XIAO ESP32-S3 board, and the README puts the parts at under 25 dollars; the vision answers come from Groq, OpenAI or a local Ollama model. What would you ask it first?

CompareeTEAM2mo agoedited

Practical notes from our verification: the MIT licence is clearly stated, but the repo itself is no longer supported — the README says the project moved to the Omi repository, so expect no fixes here. The hardware list is short and complete: a Seeed XIAO ESP32-S3 Sense, a 250 mAh EEMB battery and a printed mount (the STL is linked from the README). The single biggest time sink is not the build, it is the software: the app is an Expo project that you run on your computer and open as a local web page, with API keys for Groq and OpenAI (or a self-hosted Ollama URL) added to keys.ts. The 3D-printed mount is simple but frame-specific — test-fit before you glue or solder anything permanent. The README includes a video demo and a Discord for setup questions, but no step-by-step build video. Correction (4 October 2026): we re-checked this page line by line against the project's own repository, documentation and videos, and fixed errors in earlier versions.

BasedHardware

BasedHardware built OpenGlass for a Meta Llama 3 hackathon. The code is MIT-licensed, but the team has since moved development to its Omi repository and marks this OpenGlass repo as no longer supported.

GitHub

Star the project on GitHub

DISCLAIMER

  • Comparee is not the author of the projects featured here. All rights to each project belong to its creator — every page links to the original source, and we never host creators’ files.
  • Information is provided without warranty and may become outdated as projects evolve. Prices are indicative bands only — always check the creator’s parts list for current costs.
  • Building and operating any project is at your own responsibility. Protective equipment, safe workshop practice and compliance with local regulations are the builder’s responsibility.
GitHub repo