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
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?
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.
Category: A product, that's obviously super useful, as long as it will work.
The tech sounds cool, but my guess is this product might work for anyone whose text to command ratio is very high. If the formatting and commanding are the main tasks, then this product would be more of a distraction for a fast typist.
Regarding the website, you should lead with product, its details, rather than founders shown in 360 degrees from various vantage points. It is distracting and frankly irritating.
This is a super cool product and I am legit excited for it
But I find the idea that talking is faster than typing utterly wild, I can type faster than I can think!
Just scroll normally!
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?
This is super cool. I think it's the first piece of additional hardware I've even considered pre-ordering.
I wonder if in a couple of years we gonna have some teeth implant that will capture sounds to our devices.
When I first saw their promotional video, I thought it was a meme video with the dude mumbling to typing.
Reminds me of the Leap Motion I preordered
Fantastic!
Just preordered.
What languages do you support? (English is fine, but any others?)
Does the model require internet access, or is it running entirely locally?
God do I hate websites that make you scroll to go through an animation, why do people still do this
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.
If you're selling a physical product, using AI to model your product seems like a really bad idea. You have no idea if the product actually fits on their faces.
Calling it early - this is a scam. There's no way they're going to deliver that product, in that form factor, by Q1 2027.
Super cool! love this
Why do you hurt communicating your very cool product with an annoying and counter intuitive scroll experience?
And after some time I could not figure out what this is?
Put this part: " 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."
right at the top and maybe show someone actually doing that instead of hiding behind cryptic marketing.
Also, the product looks great.
Get your words out of the way 'say what it is'.
Very amazing tech though
Infuriating website UX, you really must add the scrollbar back. Apple do shiny things like this but they know you can't get rid of the scrollbar and piss people off trying to look at information
Q1 2027- unlikely
“Foundation models” I feel like this has lost all meaning? Can’t you just say “we train models on this data”?
Website is shit, "how it works" tells you nothing much useful about how it works, and I assume there is some custom hardware (as described in the post above) but nowhere on the website I could see this actually said, and therefore zero indication of the size of the interface electronics etc etc
How have the glasses been received? I don't understand why they're required, is it only to have a consistent mount point for the radar sensor, relative to the face? Have you thought about a throat mic?
There is a product by Subtle called Voicebuds that allows for super silent dictation. Looks like the product is not available for purchase yet, but I would prefer to have earbuds over glasses.
Let's go! Love the idea. Like the analog to bone conduction earbuds. Just please don't turn into WisprFlow where they route the voice metadata, unbeknownst to many, to a third party partner who stores and mines it for who-knows-what. Good luck fellas.
Really cool piece of tech! Languages supported rn? English only?
Preordered!
Can we use RF for other medical applications? As an anesthesiologist, we do nerve blocks before surgeries or for pain. One rough indicator that it's working is muscle paralysis. Also similar for botox.
Very interested.
This is such a weird thing that I can't wait to try it out. I don't know, like it's so crazy that it has to be good
Apple recently a purchased a company with seemingly equivalent technical capabilities:
https://www.cnet.com/tech/services-and-software/apple-qai-ac...
Hope you guys are well-positioned patent-wise.
Much respect for actual innovation. I hope you get filthy rich.
From your comments it seems to reach this point you tried at least open source lip reading software, IMUs, variant antenna mounting strategies, various frequencies, and then gathered training data. You articulated strongly why the product is desirable and why the failed modes were dead-ends. I hope you can deliver on your pre-orders and are able to scale and/or sell the IP for a strong exit.
Watched the demo video… and if this is satire: incredibly well executed
This is so dumb
There is a long way from a prototype to production. Remember Google Glasses, Oculus, Meta Rayban, Apple Vision Pro?
All of them are visionary and we would think that ppl gonna like them…
Well… ordinary people want boring tech.
This will be similar with whispering to glasses… Visionaries will like it… and return to missionary (boring) s*x (I hope that I don’t get a ban.)
Myself, I would buy these glasses immediately. But I would hate them as well as I would think I am insane (who wants to talk to yourself?).
Edit: Dang replied. I wish all the best to the team behind “whisper glasses”. It is amazing. Connect it with ElevenLabs and you rule the world. (Previous: It wasn’t meant a hate speech.)
Talk about over engineering a problem! If typing speed is an issue guess what, humans can train themselves to type really fast on a keyboard.
[dead]
FYI, the subvocal.company site is unscrollable with a mouse wheel in WebKit browsers on Mac (Safari, DDG, Kagi Orion), though the touchpad will work. No issues elsewhere (Mac non-WebKit, Win11, Ubuntu)
What's wrong with human connection and speaking. :( I feel like the isolation of this modern world is lame.
[dead]
[dead]
[dead]
[dead]
All of the images, including the glasses, are AI generated.
Anyone who preorders - you are getting scammed.
Pangram: https://www.pangram.com/history/images/614ffe91-8ba9-4d3c-b1...
EDIT: I find talking in that manner in a public setup to be awkward and unpleasant, maybe that's just me.
That said, I think the idea of having a private walkie-talkie that connects on its own frequency for large crowded spaces to be very cool
Are there any associated benchmarks with this? Like how accurate is this technology for different types of facial structures, sentences or languages?