Severe overstriding. The foot is landing far ahead of the hip on every step.
Lab-grade sprint analysis. No lab required.
Sprintalyze turns a smartphone video into a full biomechanical report in minutes. Ground contact times, joint angles, stride parameters, left–right asymmetry, and plain-language coaching insights. No markers, no hardware, no hours of post-processing.
- Body + feet at sub-pixel precision
- Minutes, not hours
- No markers. No lab.
Three steps from sprint to insight.
No markers. No motion-capture suit. Just a smartphone and a tripod.
-
01
Film
Point a smartphone at the sprinter. Record a single run at 120 fps from a tripod. No markers, no suit, no lighting setup.
120 FPS OR MORE · SIDE VIEW · 1 TAKE -
02
Analyze
A two-stage pipeline locates the athlete, then localises a sprint-optimised skeleton (full body plus heel, big toe, and small toe on each foot) with temporal smoothing tuned for ground-contact timing.
MULTI-STAGE AI · SUB-PIXEL FEET · SPRINT-TUNED -
03
Understand
Get per-step metrics, whole-stride averages, left–right asymmetry flags, and automatic coaching cues written in plain training language.
PER-STEP · ZONE-AVERAGED · INTERPRETED
Every number a sprint coach actually wants.
Not a wall of raw data. Structured metrics, per-step and per-zone, with the context needed to act on them.
-
92 msavg GCT
Ground contact & flight times
Per step, accurate to the millisecond, with inner-contact zone averaging so you measure true top-speed, not the warm-up.
-
165°peak knee drive
Joint angles
Hip, knee, and ankle across the full stride. Measured at the phases coaches actually care about: touchdown, mid-stance, toe-off.
-
4.42 Hzstep freq
Stride parameters
Step length, step frequency, and duty factor per athlete, per session, per zone. Tracked over time on athlete-level dashboards.
-
7.4 cmvertical osc.
Center-of-mass trajectory
Vertical oscillation, horizontal displacement, and leg-stiffness proxy — the signals that distinguish efficient top-speed mechanics.
-
±3.1%L/R delta
Left–right asymmetry
Flag contact-time, step-length, and joint-angle imbalances across sides. Spot hidden risk before it becomes an injury.
-
AUTOcoaching cues
Automatic interpretation
Every metric gets a band, a plain-language interpretation, and where appropriate a short coaching cue. Never prescription, always insight.
From a sprint video to a readable answer.
Two screens from the app: the full-detail Results view for when you want the raw data, and the Top Speed Feedback view for when you want a verdict.
- STRIDE LEN2.13
- STRIDE FREQ5.15
- COM DIST1.08
- GC TIME114
- FLIGHT103
- DUTY FACT0.524
Focus Areas
Flight time is a bit outside the ideal window.
Measured where it matters. Beside the track.
The honest comparison. Lab systems are still the scientific gold standard; we aim to give you 90% of that depth at 0.1% of the cost and 10% of the friction.
| Sprintalyze | Lab systems | Competitors | Visual analysis | |
|---|---|---|---|---|
| Cost | ✓ Affordable subscription | × €50k – €500k | – ~€250/yr + per-analysis | ✓ Free |
| Time to result | ✓ Minutes | × Hours to days | – Long processing | – Immediate, subjective |
| Accuracy | ✓ Multi-stage markerless AI | ✓ Gold standard | – Single-stage · MediaPipe | × Human eye |
| Usable at the track | ✓ Yes, during the session | × No | – Not practical | – Limited |
| Automatic interpretation | ✓ Plain-language cues | – Raw data only | – Raw data only | × None |
Pricing figures for other tools reflect publicly available information at time of writing. We’ll keep this comparison current and publish validation data as studies complete.
Peer-reviewed models. Practitioner-grade engineering.
The pose-estimation research under Sprintalyze will be published and open, rooted in academic gait research. We sit on top of it. Carefully.
-
Markerless pose pipeline
A multi-stage markerless AI pipeline locates the athlete and extracts a sprint-tuned skeleton: body plus detailed foot landmarks. Optimised for the joints and contact points that actually drive sprint biomechanics.
-
Signal processing tuned for sprint
Temporal filtering and outlier rejection tuned for sprint kinematics remove detection noise while preserving the high-frequency content that matters for ground-contact timing.
-
Privacy by design
Videos are processed to extract the keypoint skeleton, then deleted. Only the extracted keypoints and the athlete metadata you create are stored. Infrastructure runs on EU servers. Delete any athlete, and all their data, in one tap.
-
Validation in progress
We’re comparing Sprintalyze’s measurements against marker-based gold-standard systems in partnership with sports-science labs. Results will be published here as studies complete.
Coaches and scientists who want measured answers.
If you’ve ever said “looked better today” and wished you had a number, you’re the person we built this for.
-
Sprint coach
Track & field clubs, federation programs, collegiate teams.- Cut the “I think we looked smoother today” loop. Get measured step-by-step numbers you can track across months.
- Top-speed mechanics, acceleration mechanics, asymmetry — in a format you can show the athlete on the spot.
-
S&C coach
Football, rugby, NFL, field sports, combine prep.- Screen entire rosters at the start of pre-season without a timing system or a lab.
- Identify players leaking energy at contact, and quantify the gap instead of calling it a “feel”.
-
Sports scientist
Olympic training centers, university labs, performance diagnosticians.- Add a field-deployable markerless layer to your existing marker-based protocol.
- Access the raw keypoint data on request. We’ll publish validation as studies complete.
The questions coaches ask first.
-
Lab systems (marker-based motion capture and instrumented tracks) remain the scientific gold standard. Our internal benchmarks against marker-based systems put Sprintalyze within a few percent on the metrics that matter most for coaching — ground contact time, step length and frequency, and joint angles at touchdown and toe-off. We are running validation studies in partnership with sports-science labs and will publish results as they complete.
-
Any phone from the last few years that can record video at 120 fps or above. Most modern iPhones (8 and later) and recent flagship Android devices qualify. A simple tripod and a side-on view of the athlete is all the setup you need.
-
Videos are processed to extract a keypoint skeleton, then deleted. We only store the extracted keypoints and the athlete metadata you choose to enter. Infrastructure runs on EU servers and is GDPR-compliant. You can delete any athlete, and all their data, with a single tap.
-
You can record and queue sessions offline. The biomechanical analysis itself runs on our servers and requires a connection to upload the video for processing. Reports are cached locally once received, so you can review them again without connectivity.
-
The pipeline is tuned for sprint mechanics, which makes it directly useful for football, rugby, NFL combine prep, hockey, lacrosse, baseball base-running, and any field sport that depends on linear acceleration and top speed. Court sports and change-of-direction analysis are on the roadmap.
-
The closed beta opens in summer 2026 with a small, hand-picked group of coaches. Public launch is planned for the following season. Beta seats are prioritised for sprint and S&C coaches working with competitive athletes — sign up below to get on the list.
-
Pricing will be a monthly or annual subscription, scaled by team size. Final tiers will be announced before public launch; beta participants will receive grandfathered pricing as a thank-you for the feedback. The goal is to be cheaper than a single hour of lab time, every month.
Early access. Priority seats.
We’re onboarding a small group of coaches through summer 2026. Priority goes to sprint and S&C coaches working with competitive athletes. Leave your details and we’ll be in touch before public launch.