Skip to main content

Introduction

If you're building a platform in sports, physical therapy, performance, rehabilitation, or digital health, chances are you've needed movement data — and found that getting it requires expensive hardware, specialized setups, or custom computer vision work.

The Model Health SDK removes that barrier. Add lab-grade 3D biomechanical analysis to your app using nothing but smartphones — no motion capture suits, no ML infrastructure. Users capture video through the Model Health Companion iOS app, your platform receives clean, actionable movement data.

The SDK is currently available to select partners. Request API access to get started. The Companion iOS app is currently required for video capture. Support for additional capture methods is on the roadmap.

How It Works

The integration involves three parts working together:

  1. Your app – presents the interface, triggers recording sessions and displays data to your users.
  2. Model Health SDK – manages the recording workflow, submits data for processing in our cloud and retrieves the analysis.
  3. Model Health Companion iOS app – captures the video footage needed to run the analysis.

It typically takes an experienced developer less than a week for a complete integration — session management, calibration, recording and retrieving results. Our first pilot users were up and running in two days.

Supported Platforms

Swift SDK (iOS)

Native SDK for iOS applications.

TypeScript SDK (Web, Node.js)

Cross-platform SDK powered by WebAssembly. Works in both browser and Node.js environments.

Python SDK

Native SDK for Python applications.

Workflow Overview

A typical integration follows these steps:

  1. Session Configuration – Start and configure a data collection session
  2. Camera Calibration – Calibrate the cameras with a checkerboard pattern
  3. Subject Calibration – Record the subject standing in a neutral pose
  4. Activity Recording – Capture activities like squats, jumps, or running
  5. Activity Analysis – Retrieve recordings and processed biomechanical data

This mirrors the workflow in Model Health's own web app, so if you've used it, the concepts will feel familiar.

Support

For bug reports and feature requests, visit our GitHub repository.
For everything else, reach us at support@modelhealth.io