• 04 Jun 2021
  • 1 Minute to read
  • Dark
  • PDF


  • Dark
  • PDF

Main Features

  • Web component or iframe based integration.
  • Initiating the Rapport Cloud Processing Infrastructure (CPI) to prepare a room with the AI specified in your Rapport account and manages sessions.
  • Establishing a WebRTC connection between a browser and CPI room.
  • Handling of the streaming connection that provides audio and animation data which then gets  buffered, synchronized, and turned into a Three.js animation of your AI talking.
  • Resizing the rendering canvas to fit the parent element or set style.
  • Targeting a rendering rate of 30 fps.
  • Caching of 3D (.glb) models.
  • Sending commands to CPI modules, e.g. send-text, set-mood, set-scale, etc.
  • Receiving messages from CPI modules, e.g. for AI, TTS, ASR, AC (animation controller) modules.
  • Disconnecting sessions on page termination events where possible.
  • Info, Warning, sessionDisconnected callback and event listener notification system.
  • Providing acoustic echo cancellation (AEC) for microphone conversations where possible.

Optional Features

  • Rendering on colored or transparent canvas.
  • Providing orbital controls for allowing the user to move around the animation.
  • 3 point lighting.
  • Opening text to initiate the AI conversation on joining the room.
  • Inactivity timers for automated disconnecting sessions.
  • Displaying a loading image while a room is being prepared.
  • Callbacks and events for session connected and disconnected.
  • Configurable network signal quality indicators.
  • Configuring audio for changing audio inputs and outputs (work in progress).
  • Seamlessly switching between 'no microphone' and 'microphone required' modes.
  • Footer for displaying various status indicators.
    1. Audio controls for speaker volume and muting microphone.
    2. Status bar for displaying room preparation steps.
    3. Progress bar for displaying estimation of background work completed before joining the room.
    4. Icon showing detected emotion in user's voice.
    5. Mic icon indicates if voice activity is detected.