dynalytics
Movement is data. Patterns are prevention.
Dynalytics uses pose estimation and machine learning to analyze your climbing movement, deep analysis of movement patterns to help flag risky mechanics before they become problems. Starting with climbing, but built to adapt to any movement—other sports, manual labor, rehab, and beyond.
Who Is This For?
Climbers
Understand your movement patterns, identify weaknesses, and track your progress over time.
Physical Therapists
Access quantified movement data to better understand and treat climbing-specific injuries.
Researchers
Contribute to the growing body of knowledge in climbing biomechanics and movement analysis.
What We Track
Every frame analyzed for biomechanical insights
12 Joint Angles
Real-time tracking of shoulder, elbow, hip, and knee positioning throughout your climb
Velocity Vectors
Detailed motion analysis including speed and acceleration patterns across movement phases
Sensation Labels
Self-reported feelings of pump, fatigue, and confidence to correlate with biomechanics
How It Works
From generic knowledge to personalized insights
Base Model
Trained on many climbers
Generic Guidelines
General movement principles
You Upload Videos
Your climbing sessions
Personal Model
Learns YOUR body, YOUR style
See It In Action
Walk through the data collection experience
Upload Your Video
Simply upload a climbing video (.mov, .mp4, .avi) to begin the analysis process.
Data Pipeline
| Stage | Process | Output |
|---|---|---|
| Video Upload | Pose estimation analysis | Joint coordinates & angles |
| Motion Analysis | Velocity & acceleration calculation | Movement vectors |
| Pattern Recognition | ML model identification | Personalized insights |
Roadmap
Phase 1: MVP (Current)
Basic pose estimation and data collection platform
Phase 2: Data Collection & ML Training
Gathering labeled movement data and training the injury pattern recognition model
Phase 3: Integration & Deployment
Deploy the trained model and integrate with user-facing tools
What happens to your data?
→ Your video is processed, then deleted — we extract movement data, then the video is permanently removed when you click done.
→ Only anonymized data remains — a CSV of joint angles and positions. No images, no video, nothing identifiable.
→ Your email helps us stay in touch — we track top contributors for early access, offers, and giveaways.
Help Us Collect Data
We're building a dataset to train our pattern recognition model. Your climbing videos will help make it possible.
We'll notify you when beta access is available.