Fort, a startup that launched in January 2025, is selling a bracelet-style wearable that ditches the screen entirely and focuses on one job: logging strength workouts automatically. The device uses onboard motion and heart-rate sensors to detect exercises, count reps, and track sets along with rest periods, all without requiring the wearer to tap a display or pull out a phone. For lifters who have long relied on notebooks or clunky app interfaces between sets, the pitch is simple, but the science behind some of its more advanced claims deserves a closer look.
A Bracelet Built for the Weight Room
Fort describes itself on its official website as a “screenless wearable” built for strength training. The aluminum-bodied bracelet is waterproof and offers roughly one week of battery life, according to coverage in Wired that details the early hardware specs. Stripping out the screen is a deliberate design choice: the company argues that a display adds bulk, drains power, and tempts users to check notifications mid-set. By removing that distraction, Fort keeps the device small enough to wear during heavy barbell and dumbbell movements without interfering with grip or wrist position.
Under the hood, the bracelet packs IMU sensors, specifically an accelerometer and gyroscope, alongside PPG sensors for optical heart-rate monitoring. The IMU array handles the core workout-tracking functions. Accelerometer and gyroscope data feed algorithms that identify which exercise a user is performing, count individual repetitions, and segment each set from the rest periods in between. PPG data adds a cardiovascular layer, capturing heart rate throughout a session. The combination means the device can, in theory, produce a complete training log without any manual input.
Because there is no display, interaction happens almost entirely through the companion app. Users are expected to wear the bracelet on the same wrist consistently, then review and edit their workouts afterward if the automatic detection misses something. Fort’s bet is that removing taps and swipes during training will outweigh the occasional need to correct a mislabeled exercise later.
Preorder Campaign and Y Combinator Backing
Fort’s go-to-market path runs through Y Combinator, where the company posted a launch announcement framing the product as automated strength tracking for longevity. That page includes a discounted preorder price, positioning the device as an early-adopter opportunity before full retail availability. The longevity angle is notable because it signals Fort is not chasing competitive powerlifters or bodybuilders alone. Instead, the company appears to target a broader audience: people who lift weights primarily to maintain muscle mass, bone density, and metabolic health as they age.
The broader Y Combinator network has produced several hardware wearables over the years, though most have focused on sleep, recovery, or general activity. Fort’s narrow focus on resistance training sets it apart in that cohort. Discussion on Hacker News reflects curiosity about the device and its promise of “zero-effort” logging, but as of early 2025, no independent user reviews or third-party trial data have surfaced publicly. That leaves prospective buyers relying largely on Fort’s own claims and YC’s signal of confidence.
Velocity and Proximity to Failure
The most technically ambitious part of Fort’s value proposition is its claim to infer proximity to failure during a set. The concept works like this: as a lifter approaches muscular failure, bar velocity slows. By measuring that deceleration through its IMU sensors, the device estimates how many repetitions in reserve, or RIR, a user has left. RIR is a widely used training metric because stopping a few reps short of failure can reduce injury risk while still driving strength gains, especially for people training several times per week.
There is real science behind velocity-based estimation of RIR. A peer-reviewed study in BMC Sports Science, Medicine and Rehabilitation examined generalized velocity-based methods for gauging proximity to failure in the bench press. The researchers confirmed that velocity drops do correlate with fewer reps in reserve, giving the concept a legitimate foundation. When bar speed falls below certain thresholds, lifters are typically within a narrow band of remaining repetitions, at least under controlled conditions.
But that same paper also identified limits: the relationship between velocity and RIR is not perfectly linear, and accuracy depends heavily on the specific model used to track the decline. In other words, simply noticing that the bar is slowing down is not enough; the algorithm must be tuned to the exercise, load, and individual lifter to translate speed into a useful RIR estimate.
A separate study in PeerJ found that exercise selection and training variables significantly affect the velocity-RIR relationship. The authors reported that exercise type, training load, velocity loss threshold, and even which set in a sequence a lifter is on can all shift the curve that links bar speed to remaining reps. In practical terms, the velocity profile during a set of heavy squats looks different from one during lighter accessory work, and fatigue accumulated over multiple sets further distorts the signal.
Where the Science Gets Complicated
The challenge for Fort is not whether velocity-based RIR estimation works in controlled settings. For certain barbell lifts and well-defined protocols, it does. The challenge is whether a wrist-worn IMU can replicate the precision of the lab-grade linear position transducers and tethered encoders that most published studies rely on. A wrist sensor measures the hand’s path through space, which is only an indirect proxy for the barbell’s actual velocity. Differences in grip width, wrist angle, and individual lifting mechanics all introduce noise that can blur the connection between what the bracelet detects and what the bar is actually doing.
Research in the European Journal of Sport Science raises a more fundamental concern: mean concentric velocity may not reliably assess RIR across different contexts. The study found that targeted RIR velocity values can shift with changes in load and exercise selection, meaning a threshold calibrated for one movement may misfire on another. For a device that promises automatic exercise detection across dozens of movements, this variability is a significant technical hurdle. A model tuned for bench press at moderate loads, for example, might underestimate effort on overhead presses or overestimate it on machine-based movements.
On top of that, real-world training is messy. Lifters pause unintentionally, adjust grip mid-set, or change their tempo for reasons that have nothing to do with fatigue. Spotters may assist slightly on difficult reps, and machines can introduce friction or leverage that alter the velocity profile. All of these factors can confuse an algorithm that expects a smooth, predictable decline in speed as failure approaches.
None of this means Fort’s approach is flawed in principle. Velocity-based training is a well-established concept in sports science, and wearable IMU technology has improved substantially over the last decade. The idea of bringing that toolkit to everyday gym-goers, in a form factor that does not require attaching gadgets to bars or racks, is compelling. But the gap between lab validation on a single barbell exercise and real-world wrist-sensor accuracy across an entire gym session is wide.
At this stage, there is no publicly available independent validation of Fort’s specific algorithms. The company has not released white papers, benchmark datasets, or third-party testing results that would let coaches or researchers compare its estimates of RIR and exercise detection against gold-standard tools. Without that transparency, it is difficult to know whether Fort’s RIR numbers are precise enough to guide programming decisions or better treated as rough effort indicators.
What Early Adopters Should Expect
For prospective buyers, the most realistic expectation is that Fort will excel at basic logging before it delivers on the more ambitious promise of precise failure prediction. Automatic rep counting and set segmentation are comparatively mature use cases for IMU sensors, especially on common movements like presses, rows, and curls. Even if the device occasionally mislabels an exercise or miscounts a rep, it can still save time versus manual entry and provide a useful history of training volume over weeks and months.
The more advanced metrics, such as RIR estimates and detailed velocity curves, should be viewed as experimental until independent data suggest otherwise. Lifters who already track perceived exertion may find value in comparing their own sense of difficulty with Fort’s readings, using discrepancies as a prompt to adjust either their technique or their expectations of the device. Coaches, meanwhile, might treat Fort as an additional data point rather than a definitive arbiter of how close a set came to failure.
If Fort can demonstrate that its wrist-based measurements align reasonably well with established tools across a range of exercises and loads, it could carve out a meaningful niche in the crowded wearable market. For now, its promise sits at the intersection of solid sports science, evolving sensor technology, and a still-unproven layer of proprietary software. Early adopters drawn by the idea of a distraction-free strength tracker will be testing not just a new gadget, but a broader question: how far can a simple bracelet go in turning everyday lifting into actionable, lab grade data?
More from Morning Overview
*This article was researched with the help of AI, with human editors creating the final content.
