vitallens.js

Estimate vital signs such as heart rate, HRV, and respiratory rate from video in JavaScript.

vitallens.js is the official JavaScript client for the VitalLens API, a service for estimating physiological vital signs like heart rate, respiratory rate, and heart rate variability (HRV) from facial video.

Using a different language or platform? We also have a Python client and iOS app.

Features

Cross-Platform Compatibility:
Use vitallens.js in the browser or Node.js.
Flexible Input Support: Process video files or live streams from a webcam or any MediaStream.
Multiple Estimation Methods: Choose the method that fits your needs:
- vitallens: Provides heart rate, respiratory rate, and heart rate variability estimates. (Automatically selects the best available model for your plan. Requires an API key - get one for free on our website)
- g, chrom, pos: Offer less accurate heart rate estimates. (No API key required.)
Fast Face Detection & ROI Support:
Perform rapid face detection when required—or optionally, pass a global region of interest (ROI) to skip detection for even faster processing.
Pre-Built Web Component Widgets:
In addition to the core API, vitallens.js provides ready-to-use web components. Try our simple vitals monitor widget (as seen in the above gif), or use an advanced widget showing vitals, video, and waveforms (supports both file and webcam modes).

Installation

Node.js

Install vitallens.js via npm or yarn:

npm install vitallens
# or
yarn add vitallens

Then use it as follows:

import { VitalLens } from 'vitallens';
const vl = new VitalLens({ method: 'vitallens', apiKey: 'YOUR_API_KEY' });
const result = await vl.processVideoFile(myVideoFile);
console.log(result);

Browser

For browser usage, you have two options: using the pre-built Web Components or importing the core VitalLens class directly.

Option 1: Using the Web Components (Recommended)

This is the easiest way to get started. Just add the module script from a CDN, and you can use the custom elements directly in your HTML.

<script type="module" src="https://cdn.jsdelivr.net/npm/vitallens/dist/vitallens.browser.js"></script>

<!-- Easy-to-use vitals monitor with basic readings -->
<vitallens-vitals-monitor api-key="YOUR_API_KEY"></vitallens-vitals-monitor>

<!-- Advanced widget. Supports both file and webcam inputs, video and waveform views -->
<vitallens-widget api-key="YOUR_API_KEY"></vitallens-widget>

Option 2: Using the Core API

If you need more control, you can import the VitalLens class directly into your own JavaScript module.

<video id="my-video" autoplay muted playsinline></video>

<script type="module">
  // Import latest version from jsDelivr (recommended)
  import { VitalLens } from 'https://cdn.jsdelivr.net/npm/vitallens';

  // Or pin a specific version:
  // import { VitalLens } from 'https://cdn.jsdelivr.net/npm/[email protected]';
  
  // Or use Skypack:
  // import { VitalLens } from 'https://cdn.skypack.dev/vitallens';

  (async () => {
    try {
      const videoElement = document.getElementById('my-video');
      const stream = await navigator.mediaDevices.getUserMedia({ video: true });
      videoElement.srcObject = stream;

      const vl = new VitalLens({ method: 'vitallens', apiKey: 'YOUR_API_KEY' });

      await vl.setVideoStream(stream, videoElement); // Use await here
      vl.addEventListener('vitals', (data) => console.log(data));
      vl.startVideoStream();
    } catch (err) {
      console.error("Failed to start VitalLens:", err);
    }
  })();
</script>

How to use

Configuring `VitalLens`

When creating a new VitalLens instance, you can configure various options:

Parameter	Description	Default
`method`	Inference method: `vitallens`, `g`, `chrom`, or `pos`.	`vitallens`
`apiKey`	API key for the VitalLens API (required for method `vitallens`).	`null`
`globalRoi`	Optional region of interest for face detection (object with `{ x0, y0, x1, y1 }`).	`undefined`
`waveformMode`	Optional setting how waveform is returned: `incremental`, `windowed`, or `complete`.	(see below)
`overrideFpsTarget`	Override for the target frames per second (fps) used during inference.	`undefined`
`fDetFs`	Frequency (in Hz) at which face detection should be performed.	`1`

The default value for waveformMode is windowed if a stream is being analyzed, and complete if a file is being processed.

Methods

You can choose from several rPPG methods:

vitallens: The recommended method. Uses the VitalLens API and automatically selects the best model for your API key (e.g., VitalLens 2.0 with HRV support)
vitallens-2.0: Forces the use of the VitalLens 2.0 model.
vitallens-1.0: Forces the use of the VitalLens 1.0 model.
vitallens-1.1: Forces the use of the VitalLens 1.1 model.
pos, chrom, g: Classic rPPG algorithms that run locally and do not require an API key.

Understanding the results

When analyzing a video stream

When analyzing a video stream, VitalLens returns estimation results continuously. Each returned estimation result contains the following vital signs:

Name	Type	Based on / containing	Returned if
`ppg_waveform`	Continuous waveform	Depends on `waveformMode`	Always
`heart_rate`	Global value	Up to last 10 seconds	Face present for at least 5 seconds
`respiratory_waveform`	Continuous waveform	Depends on `waveformMode`	Using `vitallens`, `vitallens-1.0`, `vitallens-1.1`, or `vitallens-2.0`
`respiratory_rate`	Global value	Up to last 30 seconds	Face present for at least 10 seconds using `vitallens`, `vitallens-1.0`, `vitallens-1.1`, or `vitallens-2.0`
`hrv_sdnn`	Global value	Up to last 60 seconds	Face present for at least 20 seconds using `vitallens`, or `vitallens-2.0`
`hrv_rmssd`	Global value	Up to last 60 seconds	Face present for at least 20 seconds using `vitallens`, or `vitallens-2.0`
`hrv_lfhf`	Global value	Up to last 60 seconds	Face present for at least 55 seconds using `vitallens`, or `vitallens-2.0`

When analyzing a video file

When analyzing a video file, VitalLens returns one estimation result for the entire file, containing:

Name	Type	Based on / containing	Returned if
`ppg_waveform`	Continuous waveform	Depends on `waveformMode`	Always
`heart_rate`	Global value	Entire video	Video is at least 5 seconds long
`respiratory_waveform`	Continuous waveform	Depends on `waveformMode`	Using `vitallens`, `vitallens-1.0`, `vitallens-1.1`, or `vitallens-2.0`
`respiratory_rate`	Global value	Entire video	Video is at least 10 seconds long using `vitallens`, `vitallens-1.0`, `vitallens-1.1`, or `vitallens-2.0`
`hrv_sdnn`	Global value	Entire video	Face present for at least 20 seconds using `vitallens`, or `vitallens-2.0`
`hrv_rmssd`	Global value	Entire video	Face present for at least 20 seconds using `vitallens`, or `vitallens-2.0`
`hrv_lfhf`	Global value	Entire video	Face present for at least 55 seconds using `vitallens`, or `vitallens-2.0`

The library returns vital sign estimates in a structured object. vitallens.js is designed to process only a single face — so you always receive a single result object with the following structure:

export interface VitalLensResult {
  face: {
    // Detected face coordinates for each frame, formatted as [x0, y0, x1, y1].
    coordinates: Array<[number, number, number, number]>;
    // Confidence values for the face per frame.
    confidence: number[];
    // An explanatory note regarding the face detection.
    note: string;
  };
  vital_signs: {
    // Estimated global heart rate.
    heart_rate: {
      // Estimated heart rate value.
      value: number;
      // Unit of the heart rate value.
      unit: string;
      // Overall confidence of the heart rate estimation.
      confidence: number;
      // An explanatory note regarding the estimation.
      note: string;
    };
    // Other vitals...
  };
  // A list of timestamps (one per processed frame).
  time: number[];
  // The frames per second (fps) of the input video.
  fps: number;
  // The effective fps used for inference.
  estFps: number;
  // A message providing additional information about the estimation.
  message: string;
}

Core API Lifecycle: Managing the Instance

When using the core VitalLens class, you are responsible for managing the instance's lifecycle. This involves controlling the stream and listening for events.

1. Controlling the Stream

You can control a live video stream at any time using these methods:

vl.startVideoStream(): Starts or resumes processing.
vl.pauseVideoStream(): Pauses processing. The webcam stays on, but no new data is sent.
vl.stopVideoStream(): Stops processing, stops the webcam, and clears all internal buffers.

// Example: A simple pause/resume button
let isProcessing = true;
myButton.onclick = () => {
if (isProcessing) {
  vl.pauseVideoStream();
  myButton.textContent = 'Resume';
} else {
  vl.startVideoStream();
  myButton.textContent = 'Pause';
}
isProcessing = !isProcessing;
};

2. Listening for Events

Your application should listen for events to receive data and handle errors.

vitals (On Success) This event fires continuously during a stream whenever a new vital sign packet is ready.

vitallens.addEventListener('vitals', (result) => {
  console.log('Vitals:', result);
  // Update your UI here
});

streamReset (On Recoverable Error) This event fires if the network becomes too unstable. The library logs a VitalLensAPIError for debugging, stops the stream to prevent a crash, and emits this event. Your application should listen for this to handle the reset.

Best Practice: Listen for this event, notify the user, and automatically restart the stream.

vitallens.addEventListener('streamReset', (eventData) => {
  console.warn('Stream was reset:', eventData.message);
  
  // 1. Notify user and stop the old stream
  showMyErrorPopup('Connection unstable. Reconnecting...');
  vl.stopVideoStream();

  // 2. Wait 3 seconds and restart
  setTimeout(async () => {
    try {
      const stream = await navigator.mediaDevices.getUserMedia({ video: true });
      videoElement.srcObject = stream;
      await vl.setVideoStream(stream, videoElement);
      vl.startVideoStream();
      hideMyErrorPopup();
    } catch (err) {
      console.error('Failed to restart stream:', err);
      showMyErrorPopup('Could not reconnect. Please try again manually.');
    }
  }, 3000);
});

fileProgress (File Processing) This event fires multiple times when calling processVideoFile to provide text updates on the processing stages.

vitallens.addEventListener('fileProgress', (message) => {
  console.log('File progress:', message); // e.g., "Detecting faces..."
  showLoadingSpinner(message);
});

Examples

Before running any of the examples, make sure to build the project by executing:

npm run build

Also, note that each example requires an API key. Replace YOUR_API_KEY with your actual API key when running the examples.

Browser - Vitals Monitor: examples/browser/vitals_monitor.html
To run this example, execute:
```
API_KEY=YOUR_API_KEY npm run start:browser
```
Browser - Advanced Widget: examples/browser/widget.html
To run this example, execute:
```
API_KEY=YOUR_API_KEY npm run start:browser-widget
```
Browser - Minimal File Input: examples/browser/file_minimal.html
To run this example, execute:
```
API_KEY=YOUR_API_KEY npm run start:browser-file-minimal
```
Browser - Minimal Webcam Input: examples/browser/webcam_minimal.html
To run this example, execute:
```
API_KEY=YOUR_API_KEY npm run start:browser-webcam-minimal
```
Node - File Processing: examples/node/file.js
To run this example, execute:
```
API_KEY=YOUR_API_KEY npm run start:node-file
```

Try opening the HTML examples in your browser or running the Node script to see vitallens.js in action.

Troubleshooting

Error in Chrome: Refused to cross-origin redirects... This error occurs for some browsers when opening the HTML file directly from your computer, because browser security policies prevent advanced features from running in HTML files opened directly from your computer. Solution: Serve your HTML file from a local web server. In your file's directory, run:
```
npx serve
```
When to Use the Self-Contained Library Use the self-contained library if your app must run in an environment that blocks requests to public CDNs (like an offline app or behind a corporate firewall). Warning: This file is large and will significantly slow initial page load.
```
<script src="https://cdn.jsdelivr.net/npm/[email protected]/dist/vitallens.browser.selfcontained.js"></script>
```

Securing your API Key

For security reasons, we recommend that you do not expose your API key directly in client-side code. There are two primary approaches to secure your API key:

1. Run Everything on your Server

If you are building a server-side application using Node.js, your API key remains securely on your server. Simply call the API directly from your backend code without exposing your credentials.

2. Use a Proxy Server for Client-Side Code

If you need to use vitallens.js in a browser, you can set up a proxy server. The proxy server receives requests from the client, attaches your API key (stored securely on the server), and forwards the request to the VitalLens API. This way, the API key is never exposed to the client.

Our client library supports this by accepting a proxyUrl option. For example:

import { VitalLens } from 'vitallens';
const vl = new VitalLens({
  method: 'vitallens',
  proxyUrl: 'https://your-proxy-server.com/api' // URL to your deployed proxy server
});

Or when using one of our widgets:

<vitallens-widget proxy-url="https://your-proxy-server.com/api"></vitallens-widget>

Sample Proxy Server Implementation

Below is a simple Node.js/Express proxy server implementation that you can use as a starting point:

const express = require('express');
const bodyParser = require('body-parser');
const cors = require('cors');

const app = express();
const PORT = process.env.PORT || 3000;

// Securely store your API key in an environment variable
const API_KEY = process.env.VITALLENS_API_KEY;
const VITALLENS_ENDPOINT = 'https://api.rouast.com/vitallens-v3/file';

app.use(bodyParser.json({ limit: '10mb' }));

// Enable CORS for your allowed domain.
app.use(cors({
  origin: 'http://example.com', // Your allowed domain
  methods: ['GET', 'POST', 'OPTIONS'],
  allowedHeaders: ['Content-Type', 'Authorization']
}));

app.post('/', async (req, res) => {
  try {
    const response = await fetch(VITALLENS_ENDPOINT, {
      method: 'POST',
      headers: {
        'Content-Type': 'application/json',
        'x-api-key': API_KEY,
      },
      body: JSON.stringify(req.body),
    });
    const data = await response.text();
    res.status(response.status).send(data);
  } catch (error) {
    console.error('Proxy error:', error);
    res.status(500).send('Internal server error');
  }
});

app.listen(PORT, () => {
  console.log(`Proxy server listening on port ${PORT}`);
});

You can deploy this proxy server on any Node.js hosting platform (such as Heroku, Vercel, or your own server) and then set the URL as the proxyUrl in your VitalLens client configuration.

Development

Building the Library

To build the project from source, run:

npm run build

This compiles the TypeScript source and bundles the output for Node (both ESM and CommonJS), and the browser.

Running Tests

Execute the test suite with:

npm run test

For environment-specific tests, you can use:

npm run test:browser
npm run test:node
npm run test:browser-integration
npm run test:node-integration

Run specific tests:

npx jest test/core/VitalLens.browser.test.ts

Linting

Lint the code using:

npm run lint

Disclaimer

Important: vitallens.js provides vital sign estimates for general wellness purposes only. It is not intended for medical use. Always consult a healthcare professional for any medical concerns or precise clinical measurements.

Please review our Terms of Service and Privacy Policy for more details.

License

This project is licensed under the MIT License.

Name		Name	Last commit message	Last commit date
Latest commit History 328 Commits
.github		.github
__mocks__		__mocks__
assets		assets
examples		examples
models/Ultra-Light-Fast-Generic-Face-Detector-1MB		models/Ultra-Light-Fast-Generic-Face-Detector-1MB
scripts		scripts
src		src
test		test
.gitignore		.gitignore
.prettierrc		.prettierrc
LICENSE		LICENSE
README.md		README.md
eslint.config.cjs		eslint.config.cjs
jest-puppeteer.config.cjs		jest-puppeteer.config.cjs
jest.config.cjs		jest.config.cjs
jest.setup.ts		jest.setup.ts
package-lock.json		package-lock.json
package.json		package.json
rollup.config.js		rollup.config.js
tsconfig.json		tsconfig.json

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

Repository files navigation

vitallens.js

Features

Installation

Node.js

Browser

How to use

Configuring `VitalLens`

Methods

Understanding the results

When analyzing a video stream

When analyzing a video file

Core API Lifecycle: Managing the Instance

1. Controlling the Stream

2. Listening for Events

Examples

Troubleshooting

Securing your API Key

1. Run Everything on your Server

2. Use a Proxy Server for Client-Side Code

Sample Proxy Server Implementation

Development

Building the Library

Running Tests

Linting

Disclaimer

License

About

Uh oh!

Releases 6

Packages

Uh oh!

Contributors 2

Uh oh!

Languages

License

Rouast-Labs/vitallens.js

Folders and files

Latest commit

History

Repository files navigation

vitallens.js

Features

Installation

Node.js

Browser

How to use

Configuring VitalLens

Methods

Understanding the results

When analyzing a video stream

When analyzing a video file

Core API Lifecycle: Managing the Instance

1. Controlling the Stream

2. Listening for Events

Examples

Troubleshooting

Securing your API Key

1. Run Everything on your Server

2. Use a Proxy Server for Client-Side Code

Sample Proxy Server Implementation

Development

Building the Library

Running Tests

Linting

Disclaimer

License

About

Topics

Resources

License

Uh oh!

Stars

Watchers

Forks

Releases 6

Packages 0

Uh oh!

Contributors 2

Uh oh!

Languages

Configuring `VitalLens`

Packages