alt Hacker NewsShow HN: Subvocal – Talk to your computer without making noise
Hi HN, we're Aarush and Arnav from The Subvocal Company (YC F26) (https://subvocal.company/). We’re building smart glasses that let you talk to your computer without talking out loud. You can mouth words silently and discreetly and the glasses decode them to text at the speed of speech.
Here's a video of Dave from YC trying it out: https://www.youtube.com/watch?v=9QdX1eObVYI
(Edit: you don't actually have to move your lips or even open your mouth for this to work! Since the RF can see inside your mouth the system can be fine tuned on "internal speech". This isn't the automatic tongue movements that occur when "thinking" words to yourself, but more like mouthing with your lips closed. It's a bit weird to do, but with practice, we have been able to get it down pretty well.)
Arnav and I were studying CS at Georgia Tech at the beginning of the year, with no experience in silent speech or sensing. We were both AI-native builders and power users of Wispr Flow. What bugged us was not being able to use these tools in the library (or an office) or in public. At Stanford’s TreeHacks in February, we decided to build the most technically ridiculous thing we could think of. We’d read about AlterEgo, a silent speech system out of MIT’s Media Lab, and wanted to recreate it in 36 hours. We built a crude prototype, sticking EMG electrodes onto my face:
https://drive.google.com/file/d/1mwhv5HZMz3UTevdrgwS5O7g2jWV...
We got it to kind of work at a letter level in a very janky way. It wasn’t good enough to win the hackathon, but convinced us to go all in. The first time I typed a sentence without anyone noticing, I felt the magic of the technology.
We quickly realized EMG (reading muscle electrical signals) wouldn’t work:
- Signal changes due to sweat
- Electrical interference
- Placement variability
- Most importantly, no one wants to stick electrodes onto their face!!
So we switched to RF and built the glasses you see now. They emit radio waves, similar to Bluetooth, that pass through your face. As they cross the tongue, lips and soft palate, they’re modulated by silent mouthing. We train foundation models on this data to decode these perturbations to text.
Despite prior academic research, getting this to work was a huge engineering challenge. First we built an extremely low noise RF sensing stack. Then we had to place antennas so their fields overlapped maximally with the articulators. Our first attempt on glasses failed completely, so we reverted to old habits and taped antennas on our face:
https://drive.google.com/file/d/1ks0dS0eZiTV9EqSFsyB_w5BBjpc...
Since taping antennas consistently in the same spot and angle was hard, we created this weird but effective head rig to hold them in place:
https://drive.google.com/file/d/1rhzNFpdDzr7oHk0isFY70dkoV4C...
This worked, but obviously wasn’t a product yet. So we worked deep in simulations to design glasses that could hold the antennas.
That evolved into our design now, with antennas embedded within the frames. Once we established there was a signal, the next hurdle was cross-user generalization. Everyone’s facial anatomy, and hence RF signal patterns, are vastly different. We tried algorithmic workarounds but realized the only lever was to massively scale pretraining data.
So that’s what we did. We set up booths in our office and paid people to read prompts wearing our glasses, to pretrain our models on. This produced a beautiful scaling curve, with errors falling quickly as we scaled our dataset and cross-user performance becoming genuinely compelling (as you see in the video above; Dave only had a short 20-sentence calibration).
We always heard about building things from first principles but never understood it before this. We tried tackling it top down but completely failed. Only when we broke each component (the antenna, electronics, ML) down to its fundamentals to understand and optimize it did we get it to work really well.
Over the next few months we’re continuing R&D to turn our prototype glasses into the slick product on our website and ship in Q1 2027. We’re also scaling up our dataset to improve performance on edge cases.
For anyone who loves voice dictation but hates using it around people, we’ve opened preorders for our first batch of 500 units, with a $200 discount off the final price. Allocation is running out quickly, so get yours at https://subvocal.co/preorder!
Comments
Just preordered -- glasses request: there's no way to say it nicely, I have a big nose. With a high bridge. Many glasses are really tough to get properly placed; may I request you not give up on the pads w/metal adjusters that you have in the product photos? Don't let a designer take them away from you, us folks not one standard deviation from 'normal' noses will thank you.
Second, how tough is this signal processing? Will it be done on-glasses, or through an app? Third, what's the equivalent of a wake word for these? I'm reminded of the Apple Vision Pro radar to see if your fingers are clicking. I could imagine a lot of modalities where I want to just start talking and have things happen, and a lot where I don't want the glasses to pick it up. Fourth, I think a cool voice change demo would be rad -- not sure what latency is like, but if you could get it end to end under like 250ms you could probably have someone exaggerate the mouth movements and pipe out audio of them talking like someone else. Which would be another really cool demo.
Super cool. You're going to be working 80 hour weeks for the next year, I'm excited to see what you put out!
One thing you might consider is actually telling people what it is on the homepage. Took me a fair bit of work to understand that this is effectively a pair of smart glasses.
Maybe I could have figured that out from the video but for folks who can’t watch the video it’s quite hard to discover what you’re building.
I think this is incredibly cool and straight out of a sci-fi novel! I think this is phenomenal for accessibility and for hands-free computing in general. I also agree with other comments that the website should have more detail on how it works, such as what you included in the post body. Amazing work guys!
Based on my understanding of the technology, this also means you can effectively dictate in loud environments. Correct?
I find that I need this most often when I'm doing yard work, running equipment, or cleaning out the garage. Audio based microphones pick up so much noise.
If the tech is mostly in the glasses temples, I would prefer to have something that attaches to my existing glasses. I have prescription glasses and a big head, tried the Even G2 smart glasses and the available frames set just too small for me
This is:
1. Insanely cool
2. Gives me serious "whispering earring" vibes. Imagine having a full on conversation with an AI via subvocalization and listening with ear bud to guide you throughout the day. https://gwern.net/doc/fiction/science-fiction/2012-10-03-yva...
Cue the resurgence of lip reading-based espionage. Imagining having to scan the room for people watching subvocalizers. Will subvocalizers start holding up a clipboard over their face a la NFL coaches?
Do most people talk faster than they can type? It's the opposite for me - I also read aloud slower than I can read in my head.
> "you don't actually have to move your lips or even open your mouth for this to work ... It's a bit weird to do"
In the video Dave just finished the training mode and is silently moving his lips, which seems pretty natural to me. I just spent a minute trying to do 'internal speech' (as many elements of speech as possible with lips closed). It does feel weird and seems like internal speech is probably providing a fair bit less signal for the system to sense (in both signal freq and amplitude). I realized I can't guess at how much articulation is 'enough' without feedback.
Based on that I'm guessing adapting to internal speech might follow a bit different path than silent speech, which seems more like 'just do what you already do silently'. Extrapolating, I expect training internal speech might involve more real-time feedback to help the user find the right degree and kind of articulation.
I realize you're still in the prototype phase but I assume an increasing number of people are getting multiple longer sessions. Do new users advance from silent to internal speech quickly and stay there? I'd be surprised as, intuitively, 'Silent -> Internal' seems like a bigger step than 'Normal -> Silent'.
Separately, if I was using it daily, I can imagine wanting to evolve a kind of micro-expression shorthand for 'Yes', 'No', 'Pause' and 'Stop', since two humans naturally evolve similar non-verbal queues over time. Vocalizing words is flexible and expressive but not efficient. It would seem odd if an intelligent, adaptive 'invisible friend' living on my face for years didn't adapt to me and me to it. Are you guys thinking in those directions for the future?
For the ones thinking this is great, have you actually tried to talk without making sounds?
I don't think it is an intuitive action at all and I became self-conscious about not making a sound rather than talking like I normally would.
I suppose it would require some "getting used to."
Regardless, I think it is a cool idea and a real problem. Sometimes I don't want to type nor talk out loud so I find myself whispering at my phone like I'm telling it a dirty secret.
--
As I'm writing this, I think this kind of product might not hit the market for consumers, but I'm seeing this used by Flock like companies to spy on people's speech, so kudos on making machine lip reading a thing!
[delayed]
I am so glad to see an unpolished early look and an honest evaluation of about five percent error rate in the YouTube video. The really cool stuff will be the applications on top of `subvocal` though at least in my opinion.
So this could take "noise canceling" to a whole new level if you could somehow "speak" across a loud bar or across the world, for that matter. Being integrated into something like Signal or similar end to end encrypted platform would be interesting as well.
I'm more looking for the Jane from Speaker for the Dead version. In-ear device that can hear and feel the subvocalization.
I remember seeing AlterEgo a few years ago and thinking it was so cool (obviously was missing voice agents to make it really cook)
Was just thinking this could be applicable for people with some disabilities tied to an actual voice agent and speaker. Or even in foreign countries where you talk and respond in the foreign language instead of waiting for an additional translation layer.
One of the coolest things I have recently seen but would love some details. It also got me thinking a about this recent Neuralink post.
With so many people singing the praises of voice to text, I keep thinking about how most sci-fi futures feature subvocalization as a prerec for that kind of adoption. Always love to see sci-fi come to life!
You can have a silent conversation with someone else wearing the device? This does not bode well for casinos.
Very cool! I was curious to hear the details of how it worked but couldn’t find them on your landing page. I’m glad to see them here and you might want to include them more prominently on the page itself. I didn’t even realize it was a pair of glasses til I read the above.
Someone built an open-source version of this, but it’s far from polished:
https://github.com/joeblau/subvox
There was this really interesting one where it reads your lips:
https://github.com/amywork777/lipflow
I’m using these to build my own so I can work in co-working spaces, planes, and libraries. Though moving my mouth with no words might draw some eyes.
Theres a LOT of people working on this.. Cornell have been putting students on this task for about 5 years.. Lots of fun different ways of doing it. Interesting you went the radar from glasses route.. I wish you well. Yet to see a product apart from the odd lip reading ML project that makes it into a product..
Super excited to see more products in this area. I talk to my computer the majority of the time now, and I can do that because I work from home. If I were on public transportation or in an office, it would be much more difficult. It's been interesting to watch the Overton window, if you will, shift for people talking _to_ their phones. Now, the majority of my family will use voice to text on their phone whereas I remember just a few years ago, if not even just a year ago, it was made fun of.
I'm curious about whether this could be done with the hardware mounted on top of the laptop screen or on a stand on the desk, rather than the glasses form factor. Have you explored this?
I think these have a place but I can't say the same for most other new AI communication devices. Creeper glasses, pins, watches, desktop units etc. I want to be able to speak and hear from my agents at all times but why would I want a new device when I already wear earbuds? New products are being designed to pick up low murmuring/whispering, like the Subtle earbuds, but I assume most earbuds will be designed this way in future.
I knew subvocalization would be the main way to talk to ais as soon as I read it in Alistair Reynolds books. I hope it could be done without any mouth movement by just microgestures soon.
Can someone comment on the AirPods tweet making the rounds (i.e. the subvocal killer)? I am excited by Subvocal, but am wondering how much bespoke hardware is actually required?
First, I absolutely hate having to scroll and scroll and scroll and scroll to try and figure out what's going on.
Second, I read this:
"The current prototypes already decode open-ended silent speech in real time, and accuracy continues to improve as we collect more data and train larger models. Like any input method, it can occasionally get something wrong, so the product is designed to use context and easy corrections rather than pretending errors never happen."
And I have to ask, does this mean everything I vocalize is being sent to an unknown server to be analyzed, potentially read by a person, and then used to train your model further? Because that's clearly not going to fly for either me or my employer. You need to be crystal clear what information people are potentially sharing with you before they purchase your devices.
I would be fairly interested if this works without opening my mouth. Like, if you can nail _micro_ tongue movements, it would feel pretty close to thought.
I have this problem and I was wondering how it would be solved. For now I can cope because I mostly work from home but I’ll keep track of you guys. Best of luck!
This is very cool, but oddly, the site is TOO focused on the user problem being solved to the extent that it feels like snake oil. The issue is that this is contrary to people's experience of what is possible, so without discussing WHY you can do this, it looks like a scam or vaporware.
What I would really like is a washable "cyber suit". Mock or turtle neck (for the voice / STT capture / control), EEG sensors, breathing rate, blood pressure, gestures via arm motion sensors,). Removable rechargeable battery module to wash the torso / shirt slap it on a clean one. For EEG it would have to be pretty compressed though, so not sure it would be comfortable enough to wear to sleep? maybe only an internal harness needs to be that compressed. not sure. I'm tired of wearing a Fitbit, or glasses, or these different things people have to look at, keep up with, break easy etc.
When I first saw their promotional video, I thought it was a meme video with the dude mumbling to typing.
Should really just be like a leap motion device on the desk, glasses are so gimmicky and aren't gonna fly for people with prescriptions (a large set of people in-front of computers all day)
Pretty cool! My only wish is that all this kind of work didn't gravitate toward glasses or assume glasses are the default thing-on-your-face now. 3/4 of humans don't wear them, and even some of us blind bats hate them enough that we endure the hassle of contact lenses!
Neat! I don't like talking to my computer (audibly or not), but that's a cool demo.
I wonder if we could eventually end up with a "throat muscle" version where people learn to talk a new way without having to move the visible mouth muscles too.
If no one wants to dictate to their phones or computers in public, surely they don't want to do this (looking like they are talking without making any noise) in public?
Spent 30 seconds scrolling down to try to figure out what this is. Eventually clicked away from the site because it was taking too long to get to the point. I'm sure it's great and I understand it's something to do with glasses and subvocalizations, but consider a one sentence contextualization at the top of the page.
Very cool. I suspect you guys will get scooped up by a bigger player working on smart glasses soon.
This is amazing, brilliant idea and nice execution. Are you concerned about Meta or any other established smart glasses players potentially copying such an idea?
Are there any associated benchmarks with this? Like how accurate is this technology for different types of facial structures, sentences or languages?
this is super cool. I did some digging about current research on avsr using video and lip point cloud movement a while back. found projects like https://github.com/amanvirparhar/chaplin but this seems more promising!
Really neat tech! Just curious, how much did you leverage AI during the development? This seems like the type of neat tech that we should expect more and more from wielding our new tools well.
I've thought about learning sign language in order to communicate with the computer in a not-super-annoying way. Your solution... much more realistic. Cheers.
I use the Vekta. I tried whispering and it doesn't feel right.
Throat mics don't work at all - see reviews online.
This is a different approach, so I am cautiously optimistic.
Are there any health risks with this method?
Edit: mostly concerned about RF near my eyes for 8 hours a day.
That is fascinating. How do the radio waves get modulated? Do you think it is possible to use this technique for spying technology, with remote transmitters and receivers intersecting a person's facial anatomy to pick up the details of even a whispered or muttered conversation?
So it's a pair of glasses that can transcribe when you mouth words silently? I still don't know how it works, they say no cameras and it works with noise so I assume sort sort of reverse bone conducting?
Category: A product, that's obviously super useful, as long as it will work.
I wonder if in a couple of years we gonna have some teeth implant that will capture sounds to our devices.
Great work!
Really like seeing the early prototypes here, and would love to see more on how the RF tech works on the website.
Are you able to patent it?
Google wanted to put this tech into its phone, but then was convinced it was a bad idea due to privacy invasion. It could for instance correlate sexual arousal with any phone interaction (video chat, youtube, display ads, etc), giving google and advertisers the ability to make custom targeted ads for specific users, as it would "know your type".
Where does inference occur? Does it work offline?