As AI-native wearables enter the mainstream, Mentra is setting itself apart with open-source smart glasses designed for real-world use cases. From streamlining industrial assembly to enabling real-time accessibility, their platform supports a broad range of applications. I had the opportunity to sit down with Mentra’s CTO, Alexander Israelov, to learn more about the future of smart glasses, scaling AI-native hardware, and the company’s experience building on MongoDB.
How did your founder journey lead you to create Mentra?
For me, building smart glasses started around six years ago. At the time, I didn't wear glasses, but I was working on a college project late at night. I was cutting glass without safety goggles—very silly—and I got a shard lodged in my right eye.
The next morning, while the optometrist was fishing it out, I realized I’d need glasses for the first time. I put them on and immediately thought: “This is such a waste of space on our face.” Most adults in the US are wearing this tech; it has access to our eyes and ears, and we just aren’t taking advantage of it. Over the next month, I taught myself enough electronics to build smart glasses from scratch, met my co-founder on Reddit, and we met up in Vegas to show off our open-source glasses at CES (Consumer Electronics Show).
Mentra recently launched its smart glasses. What gap did you see in the ecosystem and how does Mentra Live address it?
We recently launched Mentra Live, which is our take on the Meta Ray-Bans. The Meta Ray-Bans are a great pair of smart glasses, but you can't meaningfully build for them. You can’t build a business on them like you can build an app for a phone. With Mentra OS, you can.
Our software is completely open source, allowing for massive extensibility. An exciting example is our work in manufacturing. Assembly line employees can wear our glasses so that they can look at a bolt and ask, “Hey Mentra, what's the torque spec of this bolt?” Mentra AI sees what the user sees and gives them the answer instantly, allowing the employee to know the exact specifications without breaking their workflow. It’s totally extendable; if you have a specific use case, you just plug in your data.
How did you first come across MongoDB?
We've been using MongoDB since day one. We chose it because it lets you iterate incredibly quickly. Database migrations are very easy—you can easily add columns, more data, or change structures without the friction we experienced with something like Postgres.
The main thing we like about MongoDB over Postgres is that MongoDB stores data in BSON. The way data is stored is much closer to how it's handled at runtime within the application. The conversion between the app and long-term storage is a much cleaner relationship.
The Mentra team was able to get started on MongoDB Atlas quickly, moving from signup to development in minutes without lengthy setup or procurement delays. Early on, Mentra also joined the MongoDB for Startups program, gaining access to Atlas credits, technical support, and additional resources designed to help startups build and scale faster.
How has building on MongoDB helped your team move faster and more efficiently?
The number one way is the MongoDB Atlas platform. It is a huge time saver. The profiling tools for database servers are invaluable—if a query is slow, Atlas makes it super easy to find and kill that query. We probably use Atlas every single day for debugging and profiling, as well as MongoDB’s other first-party tools.
MongoDB Compass is another amazing product. There is a feature where you can write a query for your database in natural language right in the application. Whoever thought that up should get a raise!
What parts of the Mentra stack are powered by MongoDB?
Basically every built-in app we have. We store all user settings and OS data in MongoDB. Some examples include:
- Mentra Notes: Our AI note-taker uses it to store what you see and hear.
- Live Streamer: We save all the user's platform settings (YouTube, Twitch, etc.) in MongoDB.
- Third-Party Apps: Whenever a developer builds on our platform, we give them an easy way to save and retrieve information on the backend using MongoDB.
To expand on one of those examples, the Mentra Notes app can hear what you hear and see what you see, allowing you to ask it later about things that happened during your day. So if I was in a meeting discussing a new update, I could later search for key terms related to that conversation and Mentra Notes would pull up a summary. To handle the search functionality, we have vector embeddings for every single conversation, and we use the MongoDB Voyage AI tool to recall those conversations based on the user's query.
_Spot.svg)
.svg)