Deliver Feedback Like a National Team Coach – in 60 Seconds. Forget endless hours of manual video editing. Our AI-powered Video Analysis Agent handles the heavy lifting, allowing you to focus on what you do best: developing athletes and improving technique.
Why Use the FF AI Video Agent Analyze?
- Lightning-Fast Analysis: Upload video from your mobile or GoPro. Our AI instantly identifies technical errors in classic striding, double poling, diagonal stride, and skating.
- Objective Data: Stop guessing. Get precise angles for arm pull, balance points, and push-off timing, presented with clear visual markers.
- Ready-Made Drills: The AI recommends corrective exercises tailored to each athlete’s movement patterns.
- Compare with the Best: See your athlete’s position side-by-side with World Cup pros.
“This tool saves me 70-80% of the time I used to spend on video analysis. Now all my athletes get top-tier follow-up every week.”
What the AI Detects and How It Works
The FF AI Video Analysis Agent uses computer vision trained on Nordic skiing biomechanics to identify the most common technique faults in roller ski and snow ski training. It is not a generic sports AI — it is built specifically for cross-country skiing movement patterns.
Classic Diagonal Stride Analysis
The AI measures kick timing relative to the body’s centre of mass, pole angle at plant, arm swing amplitude, and hip extension at push-off. Common faults detected include early weight transfer, insufficient kick, and insufficient hip drive. Each fault is annotated directly on the video with frame-by-frame precision.
Sample AI output – Classic Diagonal: Kick phase initiates 0.08 s early relative to optimal centre-of-mass position. Hip extension at push-off reaches 168 degrees (target: 175–180 degrees). Left arm pull terminates 12 degrees short of full range. Recommended drill: Single-leg balance kick with pole tap.
Double Poling Analysis
Double poling is the most power-dependent technique in cross-country skiing. The AI evaluates pole plant angle (optimal: 75–80 degrees from horizontal at contact), elbow position through the compression phase, trunk rotation, and finish position. Wrist and elbow angles are measured frame by frame. Athletes who plant the pole too late or shorten the pull see this quantified immediately rather than described verbally.
Sample AI output – Double Poling: Pole plant angle: 71 degrees (target 75–80 degrees). Compression phase duration: 0.31 s (slightly short). Trunk angle at compression: 42 degrees (target 40–45 degrees). Core engagement score: 7.2/10. Main improvement opportunity: plant the pole 0.05 s earlier and 4 degrees more vertical.
Skating Technique Analysis
Skating on rollerskis (V1, V2, V2 alternate) involves complex asymmetric coordination. The AI analyses push angle, glide duration per stride, arm-leg synchronisation, and lateral balance. A common finding in recreational skaters is asymmetric push — one side generating 15–25% more force than the other, which causes energy waste and cumulative knee stress. The AI flags and quantifies this asymmetry in the report.
Sample AI output – V2 Skating: Right push duration: 0.41 s, Left push duration: 0.35 s (asymmetry: 17%). Right arm pole plant 0.04 s late. Glide phase balance score: 6.4/10 (target: 8+). Recommended drill: 30-second single-skate glide balance drill, focusing on left-side extension.
Human Coach Review Statement
The FF AI Video Analysis Agent is designed to augment human coaching, not replace it. All AI output is presented as objective data: angles, timing, and asymmetry scores. Interpretation, training load management, motivational context, and injury risk assessment remain the domain of the human coach.
Our recommended workflow: the AI runs the technical measurement and generates the annotated video report. The coach reviews the flagged issues, selects the two or three highest-priority corrections for the athlete’s current training phase, and adds personalised context before sending the report to the athlete. This takes 5–10 minutes per athlete instead of 45–90 minutes of manual frame-by-frame review.
Every AI recommendation is based on biomechanical norms derived from elite Nordic skiing data. Coaches with advanced training may choose to apply different standards for developing athletes — the AI reports are always advisory, and the coaching relationship and judgment always prevail.
Camera Setup Tips for Best Results
- Side view: Best for most technique analyses — place camera 5–8 metres away at hip height. Capture at least 4–6 full strides.
- Front view: Ideal for detecting asymmetry and lateral balance. Stand 10 metres ahead of the direction of travel.
- Drone / overhead view: Best for double poling rhythm and pole symmetry. Requires drone at 8–12 metres altitude, hovering directly overhead.
Video length: 60–120 seconds is optimal. Shorter clips may not capture enough strides for statistical reliability. Longer clips are acceptable but may increase processing time.
AI Ski Coach – Technical Video Analysis
Upload a short video (max 60–120 seconds) of your skiing. Select your technique, and our AI will analyse the biomechanics — including scores, mistakes, and coaching tips.
AI Ski Coach – Technical Video Analysis
Upload a short video (max 60–120 seconds) of your skiing. Select your technique, and our AI will analyze the biomechanics – including scores, mistakes, and coaching tips.
How It Works
- Film: Capture a video of your athlete on roller skis or snow.
- Analyse: The AI scans their movements and identifies areas for improvement.
- Feedback: Send a complete report with video and drills directly to the athlete via a shareable link.
See how we measure rolling resistance.
Frequently Asked Questions
What techniques can the AI analyse?
Classic double poling, skating / freestyle, and classic diagonal stride. You select the technique when uploading your video so the AI applies the correct biomechanical model.
How long does the analysis take?
Most analyses are delivered within 60 seconds of upload. Complex or longer videos may take up to 5 minutes.
What file formats are accepted?
MP4, MOV, and WebM. Maximum file size is 150 MB. For best quality, film at 1080p / 60 fps.
Can I use footage from a GoPro or drone?
Yes. GoPro footage from a side-mounted position on a companion skier works well. Drone footage requires at least 4K resolution for accurate joint-landmark detection.
Is the AI analysis GDPR-compliant?
Yes. Uploaded videos are processed and then deleted from our servers within 30 days. We do not store personal data beyond what is necessary for the analysis report.