Sandbar is preparing to ship the Stream, a smart ring built around a single premise: the best AI gadget is a microphone on your finger. CEO Mina Fahmi demoed the device in the company's New York City offices, using it to dictate full sentences into Apple Notes by holding the ring to his mouth and pressing its touchpad. The Stream is scheduled to ship later this year and enters a category — AI-first wearables — that has so far produced more high-profile failures than shipping products.
The hardware itself is deliberately minimal. The Stream is a microphone, a battery, and a Bluetooth connection to a phone, with all transcription and model inference happening on the paired device. Fahmi's demo included a whisper mode in which he spoke so quietly that the room could not hear him, yet the transcription came back clean. A separate demo showed the ring seamlessly dictating into a Mac and then an iPhone after a firmware flash, without a visible handoff step.
“One underrated feature of the LLM revolution has been a remarkable leap in all kinds of dictation technology — even the fastest, cheapest models are getting very good at understanding and processing speech.”— David Pierce, Editor-at-Large
Sandbar's pitch is that phones are terrible input systems for the way people actually think. Unlock, open an app, tap into a text box, type — the flow adds friction to what should be a one-second thought capture. The Stream collapses that to a gesture: raise hand, press, talk. Fahmi says the idea came from his own habit of walking long distances and getting annoyed at digging out his phone every time he had an idea.
Key facts
- 01Sandbar's Stream ring is scheduled to ship later this year, led by CEO Mina Fahmi out of New York City.
- 02The device is a microphone, battery, and Bluetooth radio — dictation and AI processing happen on the paired phone or laptop.
- 03Stream demonstrated cross-device dictation, moving seamlessly from Mac to iPhone after a firmware flash.
- 04Competitors in the AI-ring category include Pebble founder Eric Migicovsky's Index 01 and meeting-focused Vocci.
- 05Established ring makers Oura and Ultrahuman are positioned to add similar voice input over time.
The ring has three defined jobs: start a chat with an AI assistant, take notes, and dictate text into any application. Most of that today runs through Sandbar's companion app, but Fahmi has been explicit that the long-term plan is to route directly into whatever assistant, notes tool, or text box the user already relies on. The cross-device dictation demo is the first visible step toward that.
Sandbar is not alone. Pebble founder Eric Migicovsky, now at Core Devices, has built the Index 01 on the same thesis — the fastest possible way to get an idea out of your head is a microphone you already have on. Vocci is pursuing a meeting-focused version. And existing ring makers Oura and Ultrahuman, with millions of units already on fingers, are positioned to add voice input as a firmware update if the category proves out.
The competitive frame matters because the rest of AI hardware has struggled. Meta is in the middle of a rough smart-glasses rollout. The Humane pin collapsed. Recording pendants from Plaud, Pocket, FoCase, and Mobvoi have found a niche among meeting-recorders and note-takers but have not broken out. The pattern in each failed launch is the same: the device tried to replace the phone.
“It's not complicated tech! It's a microphone. All the complicated stuff happens elsewhere, on the devices you almost certainly also have with you.”— David Pierce, Editor-at-Large
The ring category takes the opposite approach. Nothing about the Stream tries to obviate a smartphone or a laptop — it just moves the input surface closer to the mouth. The LLM stack does the heavy lifting on devices the user already owns, which is why the ring itself can stay small, cheap, and battery-friendly. That is a materially different bet from a $700 pin trying to be a standalone computer.
The unknowns are real. Sandbar has not published pricing, battery life, or a firm ship date beyond later this year. Cross-device dictation depends on OS-level text-injection permissions that Apple and Google both restrict, and Fahmi declined to explain how his demo worked. Whisper-mode accuracy in a controlled office demo is not the same as accuracy on a New York subway platform. And a ring with a always-on microphone raises the same social and privacy questions that dogged Google Glass, only smaller and harder to notice.
For the AI hardware market, the smart-ring thesis is the most credible answer yet to a question that has haunted the category for two years: what does an AI-native device look like when it is not competing with a phone? Sandbar, Core Devices, and Vocci are converging on the same answer — a dumb sensor on a smart network — and the incumbents at Oura and Ultrahuman have the distribution to make it real. If dictation-first rings ship well this year, the interesting AI hardware story of 2027 will not be glasses or pins. It will be the finger.
Working on something we should cover, or seeing a story we missed? Send leads, documents, or feedback to hello@aichatdaily.com. For sensitive tips, see our secure tips page for Signal and PGP options.
Spotted an error? Email hello@aichatdaily.com with the URL and the issue, or read our full corrections policy.
