The self-serve Alendis product — the app a rider or trainer uses on their own horses — works because biomechanics are cheap to produce per session. Upload a video, read the report, build a baseline. One horse at a time.
Some of the organisations that care most about equine movement do not think one horse at a time. A racing jurisdiction cares about every horse in training in its territory. A national breed programme cares about every registered animal across generations. A large training operation cares about every horse on its books, and about the operational data that links movement patterns to outcomes.
Alendis EDGE is the version of the platform built for that scale.
What EDGE is
EDGE is the same core biomechanics engine as the self-serve product, deployed differently and scoped differently. Three things change when you move from the app to EDGE.
Deployment model. The self-serve product runs in Alendis's cloud. EDGE runs on-premise or in a customer-controlled private cloud. Video never leaves the customer's infrastructure unless the customer decides it should. Data residency, retention, and access control are all configurable per deployment.
Scale of enrollment. The self-serve product asks a user to upload clips for horses they manage. EDGE ingests from the video infrastructure a customer already operates — track-side cameras, stable-camera feeds, clinic examination rigs. Every horse in the programme gets a biomechanical record continuously, without changing the workflow of anyone who works with those horses.
Audit layer. Every metric EDGE produces is traceable back to the source video frame. When a steward, veterinarian, or integrity officer reviews a flag, they can play back the exact footage that produced the measurement. This is a hard requirement in regulated environments and it shaped the platform's data model from the beginning.
What EDGE is for
Three use cases, roughly in the order we hear them from the organisations we work with.
Racing integrity and breakdown prevention
The signals that precede catastrophic breakdown in thoroughbreds — subtle stance-time asymmetry, fetlock load drift, late-phase stride compression — are measurable weeks before they become clinically obvious. In a jurisdiction where every horse's gallop work is on camera, those signals can be extracted continuously and surfaced to veterinary staff.
The value is not a single alert. The value is a per-horse trajectory that a vet can review during routine welfare inspections. A trajectory that is flat means the horse is stable. A trajectory that is drifting means "look at this one next, and here is exactly what changed." That is a meaningfully different workflow than the current pattern of periodic physical exam.
Breed programme longitudinal data
National Icelandic-horse breeding, Hanoverian studbook assessments, Thoroughbred lineage tracking — the organisations that steward breed populations have rich data on pedigree and outcomes and almost no structured data on movement. EDGE produces a per-animal biomechanical record that can be aggregated at the breed-programme level: how does a specific sire line express joint-loading patterns under work? Which conformational traits correlate with stable longitudinal baselines? The questions are old; the data to answer them has not existed at scale.
Large training operations
A thoroughbred training yard with two hundred horses in work does not have two hundred informal trainer-intuitions about each horse's movement. It has systems, and currently those systems are built around weighed work, vet-schedule throughput, and the judgement of senior staff. EDGE adds a biomechanical layer: every horse has a per-session record, the yard has a dashboard that ranks horses by drift magnitude, and the head of stables can walk into Monday morning with a ranked list of which ten horses to look at first.
How EDGE gets installed
The pattern is a six-to-eight-week deployment, broken into three phases.
Phase one: ingestion. The Alendis team works with the customer's video infrastructure team to establish a feed from existing cameras into the EDGE processing cluster. No new cabling, no trainer intervention, no visible change at the stable or track level. In most deployments, the cameras and the physical infrastructure are already producing the video; we are adding a consumer of that stream.
Phase two: baseline build. For the first four to six weeks after ingestion is live, EDGE is in observation mode. Every horse accumulates sessions against its own baseline. Staff see the dashboards but no alerts are active. This is the period where the system is being tuned to the customer's horses, surfaces, and camera geometry.
Phase three: workflow integration. Alerts are enabled in partnership with the veterinary team. Thresholds are calibrated conservatively at first — better to miss a subtle drift than to create a tidal wave of false positives that erodes trust in the system. Over the following quarter, sensitivity is tuned up as staff and veterinarians calibrate their own judgement against the platform's flags.
This is deliberately slow. Movement analysis in a regulated environment is a tool that has to earn its place, and the failure mode we worry about most is a yard or a jurisdiction rolling out the platform, getting flooded with false positives from a miscalibrated sensitivity setting, and losing trust in the data before the value is visible. The six-week tuning window is how we prevent that.
The data architecture that matters
One technical point worth being explicit about, because institutional customers ask about it first.
EDGE does not use population statistics to flag individual horses. Every metric is referenced against that horse's own longitudinal baseline. This has two consequences that matter at the programme level.
First, the platform does not create selection pressure on biomechanics. A yard using EDGE is not being told "horses with this stance-time ratio are unsuitable." The platform is telling them "this specific horse has drifted from its own history." A horse born with a three-percent stance-time asymmetry is not flagged for having that asymmetry; it is flagged if that asymmetry drifts to five percent. The programme is not biased against horses whose baseline movement is outside a population median.
Second, the data is longitudinal by construction. The value of a per-horse biomechanical history compounds. A yard that has been running EDGE for three years has a dataset that no new competitor can replicate quickly — not because of IP around the models, but because the data itself is a consequence of time under measurement. Institutional customers tend to think in decades about this. So do we.
What to do if you think EDGE is relevant
The deployment process starts with a scoping conversation — what video infrastructure exists, how many horses are in the programme, what the regulatory context is. That conversation is the cheapest part; the substantial work happens in the six-to-eight-week deployment that follows.
Institutions interested in EDGE should reach out via edge@alendis.com. We do not publish a price list for EDGE because the deployments vary substantially in scale, infrastructure, and integration depth. Every conversation starts with understanding what a given programme is trying to do.