rw-integrate-character-embedpor runwayml

Help users embed Runway Character avatar calls in React apps using the @runwayml/avatars-react SDK

npx skills add https://github.com/runwayml/skills --skill rw-integrate-character-embed

Embed Characters in React (Avatars React SDK)

PREREQUISITES:

  • +rw-check-compatibility — Project must have server-side capability (API key must never be exposed to the client)
  • +rw-fetch-api-reference — Load the latest API reference from https://docs.dev.runwayml.com/api/ before integrating
  • +rw-integrate-characters — Character (Avatar) must be created and session endpoint must exist
  • Project must use React (Next.js, Vite+React, Remix, etc.)

OPTIONAL:

  • +rw-integrate-documents — Add knowledge base before embedding

Embed real-time avatar video calls in React applications using the @runwayml/avatars-react SDK.

Installation

npm install @runwayml/avatars-react

This is a client-side package. The server-side @runwayml/sdk should already be installed from +rw-integrate-characters.

Option A: Simple — AvatarCall Component

The fastest way to embed a character. Handles WebRTC connection and renders a default UI automatically.

'use client';

import { AvatarCall } from '@runwayml/avatars-react';
import '@runwayml/avatars-react/styles.css';

export default function CharacterPage() {
  return (
    <AvatarCall
      avatarId="your-avatar-id-here"
      connectUrl="/api/avatar/session"
      onEnd={() => console.log('Call ended')}
      onError={(error) => console.error('Error:', error)}
    />
  );
}

AvatarCall Props

PropTypeDescription
avatarIdstringThe Avatar UUID from the Developer Portal or API
connectUrlstringYour server-side session endpoint (e.g., /api/avatar/session)
onEnd() => voidCalled when the call ends normally
onError(error: Error) => voidCalled on connection or runtime errors

For custom avatars created in the Developer Portal, use the Avatar UUID as avatarId.

Option B: Fully Custom — Hooks

For full control over the UI, use AvatarSession with hooks.

Components & Hooks

ExportTypeDescription
AvatarSessionComponentProvider that manages the WebRTC session
AvatarVideoComponentRenders the avatar's video stream
UserVideoComponentRenders the user's camera feed
useAvatarSessionHookAccess session state: state, sessionId, error, end()
useLocalMediaHookControl user's media: isMicEnabled, toggleMic()

Custom UI Example

'use client';

import {
  AvatarSession,
  AvatarVideo,
  UserVideo,
  useAvatarSession,
  useLocalMedia,
} from '@runwayml/avatars-react';
import type { SessionCredentials } from '@runwayml/avatars-react';

function CallUI() {
  const { state, end } = useAvatarSession();
  const { isMicEnabled, toggleMic } = useLocalMedia();

  return (
    <div className="relative w-full h-screen">
      {/* Avatar video takes full screen */}
      <AvatarVideo className="w-full h-full object-cover" />

      {/* User's camera in a small overlay */}
      <UserVideo className="absolute bottom-4 right-4 w-48 rounded-lg" />

      {/* Controls */}
      <div className="absolute bottom-4 left-4 flex gap-2">
        <button onClick={toggleMic}>
          {isMicEnabled ? 'Mute' : 'Unmute'}
        </button>
        <button onClick={end}>End Call</button>
      </div>

      {/* Connection state */}
      {state === 'connecting' && (
        <div className="absolute inset-0 flex items-center justify-center bg-black/50">
          Connecting...
        </div>
      )}
    </div>
  );
}

export function CustomAvatar({ credentials }: { credentials: SessionCredentials }) {
  return (
    <AvatarSession credentials={credentials} audio video>
      <CallUI />
    </AvatarSession>
  );
}

Fetching Credentials for Custom UI

When using the hooks approach, you need to fetch credentials from your server endpoint and pass them to AvatarSession:

'use client';

import { useState, useCallback } from 'react';
import type { SessionCredentials } from '@runwayml/avatars-react';
import { CustomAvatar } from './CustomAvatar';

export default function CharacterPage() {
  const [credentials, setCredentials] = useState<SessionCredentials | null>(null);
  const [loading, setLoading] = useState(false);

  const startCall = useCallback(async () => {
    setLoading(true);
    try {
      const res = await fetch('/api/avatar/session', {
        method: 'POST',
        headers: { 'Content-Type': 'application/json' },
        body: JSON.stringify({ avatarId: 'your-avatar-id-here' }),
      });
      const data = await res.json();
      setCredentials(data);
    } catch (error) {
      console.error('Failed to connect:', error);
    } finally {
      setLoading(false);
    }
  }, []);

  if (credentials) {
    return <CustomAvatar credentials={credentials} />;
  }

  return (
    <button onClick={startCall} disabled={loading}>
      {loading ? 'Connecting...' : 'Start Conversation'}
    </button>
  );
}

Integration Patterns

Next.js App Router (Full Example)

Server route (app/api/avatar/session/route.ts): See +rw-integrate-characters for the complete server-side session creation code.

Client page (app/character/page.tsx):

'use client';

import { AvatarCall } from '@runwayml/avatars-react';
import '@runwayml/avatars-react/styles.css';

const AVATAR_ID = process.env.NEXT_PUBLIC_AVATAR_ID || 'your-avatar-id';

export default function CharacterPage() {
  return (
    <div className="flex items-center justify-center min-h-screen">
      <AvatarCall
        avatarId={AVATAR_ID}
        connectUrl="/api/avatar/session"
        onEnd={() => window.location.reload()}
        onError={(error) => {
          console.error('Avatar error:', error);
          alert('Connection failed. Please try again.');
        }}
      />
    </div>
  );
}

Conditional Rendering (Show/Hide)

'use client';

import { useState } from 'react';
import { AvatarCall } from '@runwayml/avatars-react';
import '@runwayml/avatars-react/styles.css';

export default function SupportPage() {
  const [showAvatar, setShowAvatar] = useState(false);

  return (
    <div>
      <h1>Customer Support</h1>

      {!showAvatar ? (
        <button onClick={() => setShowAvatar(true)}>
          Talk to an Agent
        </button>
      ) : (
        <AvatarCall
          avatarId="support-agent-id"
          connectUrl="/api/avatar/session"
          onEnd={() => setShowAvatar(false)}
          onError={(error) => {
            console.error(error);
            setShowAvatar(false);
          }}
        />
      )}
    </div>
  );
}

Error Handling

Verbose Error Logging

<AvatarCall
  avatarId="your-avatar-id"
  connectUrl="/api/avatar/session"
  onError={(error) => {
    console.error('Avatar error:', error);
    console.error('Error name:', error.name);
    console.error('Error message:', error.message);
    if (error.cause) {
      console.error('Cause:', error.cause);
    }
  }}
/>

Debug Session State

import { useAvatarSession } from '@runwayml/avatars-react';

function DebugPanel() {
  const { state, sessionId, error } = useAvatarSession();

  return (
    <pre style={{ fontSize: 12, position: 'fixed', top: 0, right: 0 }}>
      {JSON.stringify({ state, sessionId, error: error?.message }, null, 2)}
    </pre>
  );
}

Browser Support

BrowserMinimum Version
Chrome74+
Firefox78+
Safari14.1+
Edge79+

Users must grant microphone permissions when prompted. Camera permissions are needed if user video is enabled.

Tips

  • Always import the styles: import '@runwayml/avatars-react/styles.css' when using AvatarCall
  • 'use client' directive is required in Next.js App Router for all components using the React SDK
  • Session max duration is 5 minutes — handle the onEnd callback to show a reconnect option
  • Credentials are one-time use — if connection fails, fetch new credentials (create a new session)
  • For the full SDK source, examples, and issue tracking: github.com/runwayml/avatars-sdk-react

NotebookLM Web Importer

Importa páginas web y videos de YouTube a NotebookLM con un clic. Utilizado por más de 200,000 usuarios.

Instalar extensión de Chrome