Meta's smart glasses are innovative, but they were missing one key feature for me: turn-by-turn navigation. Instead of waiting for a software update that might never arrive, I took matters into my own hands and built a custom turn-by-turn HUD that displays speed and directions directly in my field of view, allowing me to keep my focus on the road. The most exciting part? This was pure "vibe coding." I developed the architecture while the soccer World Cup played in the background, fueled by a serious dose of Google Antigravity. In the full write-up, I cover: 🔹 The Ingestion Layer: How I scraped Android lock-screen notifications for real-time routing data without an official API. 🔹 The Transport Layer: Why I opted for an MQTT cloud relay over local Bluetooth sockets to ensure ultra-low latency and battery life. 🔹 The Bridge: How I created a deep-link onboarding flow to get the HUD functioning smoothly inside the Meta browser. If you're interested in wearables, middleware hacking, or in seeing how we can reclaim the tech we invest in, check out the article.
As a software engineer, there are few things more satisfying than building a bridge where a vendor left a gap. I recently picked up a pair of Meta smart glasses, and while the tech is impressive, I was immediately frustrated by one glaring omission: the glasses don't navigate.
Whether you are navigating the dense, unpredictable traffic of a city commute or clutching the handlebars of a motorcycle on a winding road, the current standard for navigation is inherently flawed. Taking your eyes off the road—even for a split second—to glance down at a dashboard screen or a phone precariously mounted creates a dangerous ‘glance cost.’ You have to mentally context-switch, re-orient yourself to the map, and then re-adjust to the physical road. I wanted something better. I wanted a true, unobtrusive Heads-Up Display (HUD): a live, real-time feed of my speed and turn-by-turn directions beamed directly into my field of view.
Since Meta’s ecosystem is locked down and Google Maps doesn't exactly hand out free, real-time navigation APIs for personal projects, I had to get creative. This entire project was built with pure Google Antigravity—true vibe coding, fueled by the energy of the Soccer World Cup on in the background.
The Architecture: How It Works
The project is built around a React Native Android application that acts as a middleman between Google Maps and the glasses. It breaks down into three main layers:
1. The Ingestion Layer (Notification Scraping)
Without a direct API, the best source of real-time navigation data is the lock screen. Google Maps broadcasts highly accurate directions to the Android system notifications.
I wrote a custom NotificationModule in NativeNotificationService.ts to intercept these broadcasts. The app runs regex against the notification text, extracting instructions and distances:
To pair this with a live speedometer, the app utilizes expo-location to poll high-accuracy GPS data at a 1000ms interval:
Recommended by LinkedIn
2. The Transport Layer (MQTT amp; Ntfy.sh)
The hardest part of wearable tech is transport. Building a custom Bluetooth socket is a battery-draining nightmare. Instead, I offloaded the transport entirely to the cloud using MqttRelayService.ts.
I used Ntfy.sh, an open-source pub/sub service. The Android app takes the parsed navigation and speed data and fires off simple HTTP POST requests:
3. The Onboarding Bridge
Getting a custom web URL into the Meta glasses browser isn't intuitive. My companion app features a streamlined onboarding flow. It generates a QR code on the phone and uses a deep link to force the Meta View app to open, where you simply scan or register the HUD URL.
The Result
What started as a frustration has turned into a highly practical tool. By gluing together Android's notification listener, React Native, and an MQTT relay, I now have a live, latency-free HUD that keeps me safe and focused on the road.
Sometimes, if you want the future you were promised, you just have to build the API yourself.
Comments