Getting started

This page gets a working face enrollment running in under ten minutes. It is the only happy path; alternatives live in Guides.

Before you start, read Prerequisites and confirm you have what you need.

1. Install

The SDK lives on a private registry, so map the @sladeid scope to it once. Put this in the .npmrc next to your package.json and commit it — it holds no secret:


@sladeid:registry=https://npmjs.slade360.co.ke/

Then put your registry token in user-level config, once per machine (never in the project .npmrc — see Prerequisites for why that form fails under pnpm):


pnpm config set "//npmjs.slade360.co.ke/:_authToken" <your-token>
# or
npm config set "//npmjs.slade360.co.ke/:_authToken" <your-token>

Now install normally. No --registry flag: the scope mapping already routes @sladeid.


pnpm add @sladeid/slade-id-sdk
npm install @sladeid/slade-id-sdk
yarn add @sladeid/slade-id-sdk

Any additional runtime assets are installed as transitive dependencies.

2. Configure


import { SladeID } from '@sladeid/slade-id-sdk';
 
const sdk = new SladeID({
  middlewareUrl: 'https://id.example.com',
  auth: {
    type: 'tokenProvider',
    getToken: async () => fetchSladeIdToken(), // minted by your backend — see Concepts → Authentication
  },
});

That is the entire required config. Everything else is optional and covered in the configuration reference.

3. Mount the capture surface


<video id="camera" autoplay muted playsinline></video>
<button id="capture">Capture</button>
<pre id="status"></pre>

4. Capture and submit


import { SladeID, type FaceCaptureResult } from '@sladeid/slade-id-sdk';
 
const sdk = new SladeID({
  middlewareUrl: 'https://id.example.com',
  auth: {
    type: 'tokenProvider',
    getToken: async () => fetchSladeIdToken(), // minted by your backend — see Concepts → Authentication
  },
});
 
const video = document.querySelector<HTMLVideoElement>('#camera')!;
const status = document.querySelector<HTMLPreElement>('#status')!;
const button = document.querySelector<HTMLButtonElement>('#capture')!;
 
const session = sdk.createFaceSession(video, {
  autoCapture: true,
});
 
session.on('guidance', (g) => {
  status.textContent = `phase: ${g.phase}`;
});
 
session.on('captured', async ({ result }: { result: FaceCaptureResult }) => {
  try {
    const enrolled = await sdk.enrollFace('demo-subject-001', result);
    status.textContent = `enrolled id=${enrolled.id}`;
  } catch (err) {
    status.textContent = `error: ${(err as Error).message}`;
  }
});
 
button.addEventListener('click', () => session.start());

Expected response shape from enrollFace:


{
  "id": "f3c4b2a8-...",
  "enrollee": "demo-subject-001",
  "createdAt": "2026-05-14T10:00:00Z",
  "raw": { "...": "untouched backend payload" }
}

What just happened

SladeID wired up the authenticated transport and the biometrics API for you.
createFaceSession opened the camera, loaded the face landmarker, and started a per-frame guidance loop. See Capture session lifecycle.
The 'captured' event fired once the face passed the quality gate. See Quality scoring.
enrollFace submitted the capture to the Slade ID backend over HTTPS with your token.