Startup Sabi is building a beanie that reads your thoughts and turns them into text
California startup Sabi is developing a wearable device shaped like an ordinary beanie. It reads the brain's electrical activity and turns thoughts into text in real time. The device relies on non-invasive EEG sensors, so it requires no surgery and no implanted electrodes. The founders say this is a first step toward an affordable, mass-market brain-computer interface.
AI-processed from Wired; edited by Hamidun News
A hat that reads your mind sounds like science fiction, but that is exactly what California startup Sabi is promising. The company is developing a wearable device shaped like an ordinary knit beanie. Its built-in sensors read the brain's electrical activity and convert it into text in real time.
If the technology works as intended, it would be one of the first real steps toward what researchers and writers have long called the cyborg era, when the line between human and machine starts to blur not in an operating room but in your closet. Brain-computer interfaces have existed for decades. But they have so far stayed confined to clinics, military labs, and highly specialized medical devices.
The main reason is complexity and invasiveness: the most accurate BCI systems require surgically implanting electrodes directly into the cerebral cortex. That is the path Neuralink has taken, and its first clinical results are genuinely impressive: paralyzed patients have controlled a computer with their thoughts. But you can't build a mass market on surgical implants.
Sabi is betting on the opposite approach: a fully non-invasive interface built into a piece of clothing that you put on in the morning like anything else. The device is based on electroencephalography. EEG sensors built into the beanie pick up the faint electrical signals the brain generates when a person intends to say a word or mentally rehearses the movements of speech, known as imagined speech.
Machine learning algorithms process these signals, recognize characteristic patterns, and decode what the user intends to say. The result is text that appears on screen without a single keystroke and without a word spoken aloud. The key technical problem for non-invasive EEG devices is the signal-to-noise ratio.
Brain signals picked up from the scalp are roughly a thousand times weaker than those from implanted electrodes, and they are easily drowned out by artifacts: muscle movement, blinking, external electromagnetic interference. That is why most attempts to build a mind reader in the form factor of a consumer headset or hat have produced limited results: the devices either worked with only a narrow set of commands or required hours of individual calibration. A breakthrough is only possible with more powerful algorithms, and this is where modern neural networks open up new possibilities.
Sabi has not yet disclosed technical details about its algorithms and has not published specific recognition accuracy figures. That is typical for early-stage startups, for which attracting attention and investment matters more than answering hard technical questions. Wired describes the device as one that could open the door to a cyborg future, a framing that for now is closer to a marketing narrative than a verified fact.
Still, the direction is strategically important. The wearable neural interface industry is clearly on the rise: Meta is investing in reading speech with EMG wristbands, Emotiv and Neurosity sell EEG headsets for developers, and OpenBCI is building an open ecosystem for neurohackers. What fundamentally sets Sabi apart is its focus on mass-market consumers, with a product that doesn't look like medical or tech equipment.
Just a hat. Nothing extra on the outside. If the company manages to achieve acceptable accuracy and reliability in real-world conditions, the consequences will be significant.
First and foremost for people with speech impairments and motor disorders: Parkinson's disease, amyotrophic lateral sclerosis, and the aftermath of stroke. For them, an affordable non-invasive brain-computer interface could become a fundamentally new way to communicate with the world. The next audience is the productivity market: typing via imagined speech is, in theory, faster than any keyboard.
For now, Sabi is a bet on a technology that has not yet proven itself in mass use. But it's exactly these kinds of bets — on non-invasiveness, an everyday form factor, and consumer scale — that set the next wave of neurotech apart from the last one. A hat that reads your mind is getting closer and closer to reality.
*Meta is designated an extremist organization and banned in the Russian Federation.
- Nov 1, 2026Local AI models agreed to pay scammers, while simple rules blocked all 42 schemes
- Oct 31, 2026Open-source LLM gateway KTMS 1.0 plugs your own GPUs into the cloud and survives outages
- Oct 31, 2026A 22B-Parameter Video Model Runs on an RTX 3080 with 10 GB of VRAM
- Oct 29, 2026T-Tech Introduces LIMe, a Method That Trains Language Models With 15% Less Compute
Build & Ship Production AI Agents Without Boilerplate Code
Start your free 72-hour trial at Hamidun Academy (academy.hamidun.com) — interactive lessons on Claude Code & Cursor, live Tech Radar, and production agent templates.
The AI world, distilled — once a week
Seven stories that actually mattered, hand-picked. No noise, no reposts, no press releases.
Done! Check your inbox for a confirmation.