Your annual physical is fifteen minutes. Your body runs the other 8,760 hours a year alone. Prana is building an AI that watches the gap - and pulls in a human when it matters.
The pitch fits in one uncomfortable sentence: your Oura ring has never spoken to your Epic chart. Prana, a San Francisco company out of Y Combinator's Winter 2026 batch, treats that silence as the actual disease. Not the high blood pressure or the creeping glucose - those are symptoms. The disease is that nobody, and nothing, is watching the space between your appointments.
Primary care runs on a strange schedule. You see a doctor for fifteen minutes once a year, answer a few questions, get some bloodwork, and then vanish back into your life for another twelve months. Prana's founders call the thing that happens in that stretch "clinical drift" - the slow, unremarkable slide most people never notice and most physicians never see, because they only get one window a year to look through.
Prana's answer is to fill the gap with software that does not sleep. It connects to your medical records and your wearables, folds both into a single clinical timeline, and runs an AI over that stream around the clock. When the numbers start drifting - a resting heart rate that keeps ticking up, glucose readings sliding the wrong way - the system flags it and, importantly, escalates to a real physician. The company describes it as a 24/7 intelligent layer sitting on top of your health data. The unofficial version, from the founders themselves: an AI doctor in your pocket.
Two problems stack on top of each other. The first is time. A once-a-year physical is a snapshot, and snapshots miss motion. The second is fragmentation. Your health information lives in scattered silos - a hospital's electronic record here, a wearable's app there, a lab result buried in an email - and none of them share context. A doctor reading your chart at an annual visit is working with a partial picture assembled a year too late.
The result, Prana argues, is late detection - metabolic disease and chronic illness caught after they have already dug in, rather than in the weeks when they were still just a trend line bending the wrong way. That framing is what turns Prana from "another telehealth app" into something with a clearer thesis: the value is not the visit, it is the watching.
It is worth sitting with how ordinary the failure is. Nobody is being negligent. The doctor is doing exactly what the fifteen-minute format allows, and the patient is doing exactly what a busy life allows, which is nothing until something hurts. The gap is structural, not personal. Prana's founders seem to have started from that observation rather than from a piece of technology looking for a use - the product is shaped around a scheduling problem in medicine, and the AI is the tool that happens to fit it, not the reason the company exists.
Under the hood, Prana is a set of integrations stitched into a monitoring engine. It ingests medical records through a records-aggregation partner and pulls in real-time biometrics from wearables - Fitbit, Oura, Garmin, Cronometer and the like - through a data-integration partner. Those two streams become one timeline. From there, the AI does the continuous reading, and the human network does the deciding.
Import records from 25,000+ providers and sync your wearables.
History and live biometrics merge into one clinical timeline.
AI watches heart rate, sleep, activity and labs for drift and anomalies.
Real risks route to a physician for chat, prescriptions, or lab orders.
The consumer-facing surface is deliberately plain. You can open the app, chat with the AI about a symptom or a confusing lab result for free, and if you need a human, book a physician who can send a prescription to your pharmacy or order deeper bloodwork. It is available now as a live beta, including an iOS app. The heavier promise - the continuous, escalating monitoring - is the part the founders are betting the company on.
Prana priced itself like a consumer product rather than an insurance claim, and the structure is the argument. Free AI chat is the front door. A visit with a real doctor is a flat fee, no insurance card required, no subscription wall. For people who want the always-on monitoring, there is a paid tier. It is a funnel that most incumbents in American healthcare structurally cannot copy without unwinding how they bill.
| Tier | Price | What you get |
|---|---|---|
| AI Doctor chat | Free | 24/7 AI health assistant - questions, symptoms, lab explanations |
| Physician visit | $39 | Flat fee, no insurance, no subscription; prescriptions and lab orders |
| Prana Health Pro | $19.99/mo | Continuous monitoring and advanced features; 7-day free trial |
There is a second customer hiding behind the consumer one: self-insured employers. Companies that pay their own health claims have a direct financial stake in catching metabolic risk early, and Prana has been aiming its monitoring pitch at exactly that audience. It is a familiar pattern in digital health - win the individual with a clean app, sell the aggregate to the people who foot the bill. The founders' public asks have leaned that direction too: trial signups, and introductions to benefits leaders at companies that carry their own risk.
The pricing also does something quieter. By making the AI chat free and the human visit cheap and flat, Prana lowers the cost of asking a small question - the kind of "is this worth worrying about?" query that most people currently either google at 2am or ignore entirely. Lower that friction enough and you change behavior, which is the whole game in preventive care. The revenue may come from the subscription and the employer contracts, but the moat, if there is one, is in becoming the thing people actually open when they are unsure.
The team is small and pointed. Meer Patel, the CEO, studied biomedical engineering at Johns Hopkins and had a seat waiting at Brown Medical School. He deferred it to build Prana, on the theory quoted above - that software reaches more people than a full appointment book ever could. Before Prana he worked as an AI engineer scaling generative systems for large enterprise clients.
Vishvam Rawal, the CTO, comes from quantitative finance, with time spent around firms like Barclays before pointing that machinery at a different kind of ticker. High-frequency trading is, at bottom, the art of reading fast-moving signal streams and reacting before everyone else - a description that maps almost too neatly onto continuous health monitoring. Sanjit Menon, the clinical lead, is a graduating MD who previously co-founded and scaled an AI medical-education platform to hundreds of customers. Between them they cover the three surfaces Prana has to get right at once: the data plumbing, the algorithms, and the medicine.
Prana lands in a crowded but splintered neighborhood. On one side sit the proactive-testing brands that sell you a big annual panel and a dashboard. On the other sit the telehealth services that connect you to a doctor on demand but forget you the moment the call ends. Prana's wager is that the interesting product is the bridge between them: not a one-time test and not a one-off visit, but a continuous line with a human standing by when the line bends.
There is also the question of trust, which in health moves slower than in most software. People will hand a chatbot their vacation plans without a thought and hesitate before handing it a decade of lab results. Prana's doctor-in-the-loop design is partly an answer to that - the AI never quietly makes the call on its own; a human is the one who writes the prescription or orders the panel. That structure keeps a licensed professional accountable at the moments that matter, which is both a safety feature and, frankly, a requirement for anyone who wants to touch real medicine rather than wellness advice.
Whether that holds depends on the boring, hard things - keeping integrations alive across thousands of record systems, earning trust with medical data, proving the monitoring actually catches problems earlier than a yearly visit would. None of that is guaranteed. But the framing is unusually clean for a health startup: sell the watching, not the appointment. In a system built entirely around appointments, that is either a small idea or a very large one.