Haptic
Market — a quieter
way to trade.

Haptic Market is a smart ring that pairs with a small fleet of AI agents to watch the stock market on your behalf — and only pulse when there is a trade worth your attention. The phone is no longer in the loop.

The ring sits on the index finger and is paired with a phone-side service that runs a personal portfolio of agents — one watching price action, one watching news, one watching the user's own historical biases. The agents argue with each other in the background. When they reach a confident, validated recommendation, the ring delivers it as a short haptic pattern. A tiny OLED screen built into the bezel surfaces the rest: ticker, direction, magnitude, and the agent's reasoning in a single line.

The product is a deliberate response to financial doomscrolling — the modern condition of opening Robinhood, scrolling charts, refreshing portfolios, and walking away more anxious and more wrong than before.

Modern retail-trading apps — Robinhood, Webull, the rest — solved the access problem and accidentally created two new ones. For the cautious newcomer, the firehose of charts, options chains, and breaking-news pings is paralyzing; for the freshly confident novice, the same firehose feels like information, and produces the wrong kind of courage. Either way, the user ends up in the app instead of away from it.

The pattern is identifiable across user research: open the app to check a position, scroll a chart, doomscroll a feed, place a trade that was not on the morning's plan, regret it before the day closes. The interface is doing exactly what it was designed to do — surface activity — and the activity is the harm.

The opportunity, I felt, was to invert the relationship: let the data come to the user only when it matters, through a channel that physically cannot hold their attention for very long.

The system rests on three pieces. Three AI agents, running on the phone, each with a single job: one watches price action, one watches the news graph, one watches the user's own past mistakes. The agents must agree — or at least cite each other — before a recommendation surfaces. If a math signal is bullish but the news graph is contradicting it, the alert is held back. Oversight is the product.

Surfaced alerts arrive as haptic patterns first. I wrote a small dictionary of pulses: one short pulse for an informational nudge, three measured pulses for an actionable buy, two-and-a-tail for an exit, a long sustained vibration for a position the agents believe the user is about to make in error. The grammar is intentionally small — the wrist learns it in a day.

When the user lifts their hand to look, the bezel OLED — a roughly 19 mm × 8 mm screen — surfaces a single, single-line briefing: ticker, direction, magnitude, the loudest reason, and a confidence score. To accept, the user double-taps the bezel. To dismiss, they do nothing. There is no chart. There is no feed.

The prototype is a web simulation of a single trading day. The user wears the ring (rendered on-screen and pulsing via the device's vibration API on mobile), receives three alerts across the day, and watches the bezel surface unfold. Every alert can be opened, accepted, dismissed, or — most importantly — refused on the user's own past behavior by the bias agent.

What I learned, and what surprised me, is that removing the chart was the move users responded to most. The chart had become a comfort object: it didn't help, but its absence felt like a loss. Replacing it with a single sentence of reasoning created visible relief in the testing sessions — people looked up from the device sooner, talked about positions less, and reported feeling less “watched” by the market.

The next iteration tightens the bias agent — the hardest of the three to train, and the one with the most personal-data sensitivity. The long-term ambition is a piece of jewelry that is not, in any meaningful sense, a screen. Just a small, considered partner that keeps you honest about money.