Getting Started — Your First Talking AI Avatar

We're going to take you from zero to a working, talking AI avatar in about five minutes, once you have a Kaltura account — copy and paste the commands as we go.

On Windows? Run these commands in WSL2, Git Bash, or PowerShell — every command below is a plain node/npm call with no Bash-only syntax, so it works as-is in PowerShell too.

What you need before you start

1. A Kaltura account with the Agentic Avatar feature enabled. No account yet? Start a free trial.
2. Two pieces of information from your account — we'll get them in Step 1:
   • your Partner ID (a number, like 1234567)
   • your Admin Secret (a long string of letters and numbers)
3. A computer with a terminal and Node.js 18+ installed. Check with node --version; install from nodejs.org if it's missing.

That's it — we don't need to install anything from Kaltura.

Step 1 — Get your credentials

Our Partner ID and Admin Secret are our keys to the system.

1. Log in to the Kaltura Rich Media CMS at kmc.kaltura.com (or your account's custom URL, if your organization has one).
2. Go to Settings → Integration Settings.
3. Copy the Partner ID (shown at the top) and the Administrator Secret.

Keep the Admin Secret private. It's like a password — never paste it into a public chat, a screenshot, or a file you might share.

Step 2 — Set up the project

We'll clone the repository and install the quickstart's dependencies:

git clone https://github.com/kaltura/intelligent-agents-sdk.git
cd intelligent-agents-sdk/quickstart
npm install

Then we'll set our credentials for this shell session:

export AGENTIC_PARTNER_ID=1234567
export AGENTIC_ADMIN_SECRET=your_admin_secret_here

On PowerShell? Use $env:AGENTIC_PARTNER_ID="1234567" instead of export.

Step 3 — Create our own agent from scratch

This one command builds a brand-new agent — brain, face, voice, and all — from a plain-English description:

node create-agent.mjs "A friendly yoga studio receptionist who helps people book classes and answers questions about memberships"

We'll see progress messages as it builds the brain, face, and voice. At the end it sends a smoke-test message and prints the reply, plus the new IDs (configId, agentId, avatarId, widgetId) we need to embed or extend the agent.

Step 4 — Talk to our agent again

create-agent.mjs already sent one smoke-test message for us. To send more, we'll use the SDK's headless converseOnce() directly — save this as a small script:

import { Management } from '@kaltura/intelligent-agents/management';

const kaltura = new Management({
  partnerId: process.env.AGENTIC_PARTNER_ID,
  adminSecret: process.env.AGENTIC_ADMIN_SECRET,
});

const reply = await kaltura.converseOnce('<configId from Step 3>', 'Hello! What can you help me with?');
console.log(reply.text);

converseOnce() mints its own conversation token from the configId — our admin secret never leaves the process.

What we just did

Where to go next