alex1115alex

The QUERY_ALL_PACKAGES permission (what an Android app needs to see all the packages installed on your phone) is a little weird. The user doesn’t get prompted and explicitly grant permission for it like they would for something like MICROPHONE- having it in the app’s manifest alone is sufficient to query packages. However, Google Play Console does make you submit a video of how the permission is used in your app in order to publish on Google Play if they detect it in your manifest.

The acceptance criteria made sense for our app (it displays your phone’s notifications on your smart glasses HUD, and users need a way of selecting which apps can/can’t display notifications). I don’t know how HSBC justifies it though.

We're going to be putting out a Mentra Edge SDK in the next few months, but it comes with some downsides. Using your phone as a compute device is a battery hog, and you can only connect one app to the glasses at a time.

It has always been possible, we've even seen someone in our Discord host it directly on their phone using Termux:

https://github.com/Mentra-Community/MentraOS/tree/main/cloud

Hey thanks for commenting. Developers host their own apps - they don't run on our servers. See:

https://docs.mentra.glass/ubuntu-deployment

https://docs.mentra.glass/railway-deployment

In your opinion, what do you think should change to make this an "actual" open source OS?

Mentra Live runs AOSP similar to the other AI glasses on the market (Ray-Ban, Xiaomi AI Glasses, RayNeo V3 AI Glasses, etc). It's heavy, but allows us to ship fast. You'll find this code in `asg_client` folder.

We're also working on a pair of HUD glasses that will release in 2026 using an NRF5340 MCU. The code for this is being developed in the `mcu_client` folder.

MentraOS is a cloud OS (there doesn't exist a good way to build this on the edge, we've tried). Users don't need our cloud service, however, as it can be self-hosted:

https://github.com/Mentra-Community/MentraOS/tree/main/cloud

https://cloud-docs.mentra.glass/

It's all open-source:

https://github.com/Mentra-Community/MentraOS

Hey, thanks for commenting, but this isn't true. You can check out our glasses compatibility list here:

https://github.com/Mentra-Community/MentraOS/blob/main/glass...

We're also looking to support the Brilliant Labs Halo glasses once they release later this year.

Regarding Ray-Bans: We'd love to support those, but the Ray-Bans are extremely locked down. Nobody has found a jailbreak yet. We're always open to support more glasses provided they're all-day wearable and have an SDK.

In terms of all day wearable HUD glasses:

Even Realities G1 are the best HUD glasses on the market right now. They’re the first pair (with prescription) that I can wear all-day without pain, and without looking like a dork.

My team used to main the Vuzix Z100 glasses, starting with the Vuzix Ultralite reference design that predated them. We won’t touch them these days (and recently stopped selling them on our store).

Others… Meizu StarV Air 2 and INMO GO2: both lack public SDK, GO2 is too heavy. Brilliant Labs Frame: cool prototyping toy, awful glasses.

For “AI glasses” that have camera, no display:

You have the RayBan and a number of companies making these. The only one I can recommend is our upcoming Mentra Live (https://mentra.glass). It has the same camera sensor as the RayBan, but runs open source software & has an SDK.

For more sci-fi glasses that run Android and have display + camera, see the INMO Air 3, TCL RayNeo X3. These are too heavy to be worn as regular glasses, but are fun prototyping tools.

All these companies will exist in 2026. As for a 5 year horizon, I’d place my bets on Even Realities, and Vuzix (as a waveguide supplier, not consumer HW). Meizu and TCL will stick around as Chinese megacorps, though I’m 50/50 they will continue developing consumer smart glasses. Brilliant Labs is cooked unless they turn things around with their next pair of glasses.

Google & Android XR: I don’t expect their glasses to be competitive for at least a few HW generations at minimum. In terms of public information, we know they’re monocular and heavy (>45g), which is an immediate killer for the majority of users.

I have a pair - they're not as integrated as you'd think. It's essentially a BLE device that projects text/data sent from your phone, so any data transmission depends on the companion app you use.

Shameless plug: We build an open source OS for glasses that works with them. AugmentOS.org

This exists, sort of. I got a demo of these at a YC event a couple months ago and was very impressed. The back is opaque, though, so it's not real AR, but it does accomplish "E-Ink displays in a glasses form factor":

https://solreader.com/

You don’t have to be incorporated for Azure startup program. Sole proprietor LLC works fine. It’s kind of an insane deal… $125k credits for a $150 LLC and a YouTube demo.

The Unitree robots look great in videos, kind of suck in reality. The motors in the humanoid (G1) overheat after shaking hands a few times, and the wheeled dog (GO2W) drifts like a broken RC car and constantly topples over in motion.

They also patched the known jailbreak methods early this year, so all newer models lack sensor access unless you pay Unitree massive $$$ for SDK access.

The base Go2 is a fun toy, though. There’s a high level web SDK you can use for free.

You can absolutely find more feature-packed glasses. Epson, XReal, Viture, Vuzix Blade, TCL RayNeo X2/X3, INMO Air 1-3, etc. The issue is all these weigh north of 70g (too heavy for comfortable daily wear), and/or require a wire that connects to an external compute box (not socially acceptable for daily wear). Glasses like the Even Realities, otoh, are lightweight, look normal, and are wireless, so can be worn all-day as you would normal prescription glasses.

Mentra here.

The Nordic MCU they use isn't actually the limiting factor, rather it's the glasses' firmware. For bitmaps from third party apps (like AugmentOS), they enforce 194 byte chunk sizes and do not support RLE. Their first-party app does not have these limitations. We're stuck with this problem for the G1, but we're working with hardware partners to make sure future glasses don't have these issues.

The hotspot settings are particularly barren. You can turn the hotspot on/off, change the password... and... that's it. They can't be used for anything useful, like detecting when a device connects/disconnects. I'm not sure why they added Shortcuts in this state.

Our team is filled with language learning nerds. Something we found while learning Chinese and being in China is that one of the hardest parts of language learning are the uncommon words.

For example, when a native speaker says something like:

"The manufacturing happens on 6th floor",

We'll know the words for

"The [unknown] happens on the 6th floor",

but not the word for "manufacturing".

Our strategy is to fill these gaps, and only these gaps. This way, we can learn uncommon words faster over time without having to rely on full-on translation as a crutch.

AugmentOS is open source, so feel free to self-host, or even reference our communicator code in our repo to interact with the glasses directly via BLE.

This is an inferior means of development, however.

By going through AugmentOS, you get a much easier development experience, compatibility with multiple pairs of glasses (through both iOS and Android), and the ability to run multiple glasses apps simultaneously.

You can get prescription inserts from Vuzix, but they're pretty bad. If you need a prescription, and want to run AugmentOS, your best bet is to buy the Even Realities G1 instead.

One of our users documented projecting sheet music onto his smart glasses's display (with a HUD). He did a great job documenting the limitations of 2025's tech, but it gives a great look into what's going to be feasible next year.

Awesome job Kevin!

Surprised to see so much love for Kate here. I've always felt it has too much UI clutter for to make for a good Notepad.exe alternative, but too few features to be a good VSCode alternative. It seems most default KDE apps suffer from issues like this, to the point where I ended up using for VSCode for everything on my Linux machine (before eventually switching to MacOS and just using the Notes app for everything that isn't code).

One of our app devs built this recently, but for swearing:

https://youtube.com/shorts/FthRCwn1JuM?si=lC3eWAUI7sV-LL-r

A wearable speech coach would be awesome, though. Detect filler words and give you an alert on your HUD when it detects "uh" "uhm" etc.

Preach. My team's building an OS for smart glasses and some of our most common feature requests are iOS notifications & being able to reply to them.

We're going to have to do insane things to get them working. Due to how ANCS works, we're considering developing an ANCS "doohicky" (either a BLE pop-socket, smart-ring, or mag-safe wallet) which gets notifications via BLE & relays them back to the iPhone, to then send to the glasses. That would just get us the raw notifications, though, and wouldn't solve the issue of replying. The other option is a Beeper-like system in the cloud to bypass iOS entirely, but that also has downsides.

It's a total mess, especially compared to Android where you can just easily listen for notifications & send them to the glasses without much pushback from the system.

Awesome to see Firebender launch here! My team discovered it a month ago after ages of trying (and failing) to find a good "Cursor but for Android Studio" and we've been using it daily ever since. Yesterday, with one prompt, I was able to use the "agent" feature to build an activity that outputs to the display of monocular smart glasses. WITH ONE PROMPT!

Also, the developers are super active in their Discord and are fast to fix bugs. Love this project.

Awesome job launching guys! We used Browser Use last week to order burgers from our smart glasses:

https://x.com/caydengineer/status/1889835639316807980

One thing I'm hoping for is an increase in speed. Right now, the agent is slow for complex tasks, so we're still in an era where it might be better to codify popular tasks (eg: sending a WhatsApp message) instead of handling them with browser automation. Have yall looked into Groq / Cerberus?

Good question! Convoscope already has an Android app. While its primary use-case is displaying live results on smart glasses, it can also be used standalone with results displaying on-screen, or read aloud with TTS. This is also the premise of our web frontend.