Real-time HD video, audio, and screen sharing powered by LiveKit. Call or hail a colleague, share an AI-watched screen, and read live captions with on-the-fly translation — all without leaving your organization.
Your Office AI uses a few different kinds of real-time room, each for a different purpose.
| Room | What it's for | Who's in it |
|---|---|---|
| Organization presence | Ambient presence — see who's around. Auto-connects when an organization is active. | Members of the organization |
| Solo-AI room | Share your screen and mic with the AI only; the AI can see a preview of your screen. | Just you and the assistant |
| Workspace room | Direct calls and hails within a workspace. | Workspace members you invite |
Within a workspace room, Call and Hail use the same underlying room but start interaction differently:
You can multi-select participants in the grid and then batch Call or Hail them at once. Pin a participant to keep them front and center.
Use Hail when you're not sure someone is free — it's a gentle ping. Use Call when you've already agreed to talk and want to jump straight in.
Go to the workspace whose members you want to talk to.
Select one or more participants in the grid, then Call to join directly or Hail to ping them first.
Pick your camera and microphone, then join. You can turn your camera off to go audio-only.
Share your screen, and toggle live captions — with translation — from the call controls.
The call toolbar gives you control over your experience.
| Control | What it does |
|---|---|
| 🎤 Mute / Unmute | Toggle your microphone. |
| 📷 Camera On / Off | Start or stop your video; turn it off to go audio-only. |
| 🖥 Screen Share | Share your entire screen, a window, or a browser tab. |
| 🔤 Live Captions | Turn captions on; choose plain transcript or translated text. |
| 📌 Pin | Keep a chosen participant front and center in the grid. |
| 🚪 Leave | Exit the call; others stay connected. |
Your Office AI shows connection-quality stats for the call, alongside a "watching you" indicator that shows who is currently subscribed to your video.
Share your content with one click. Your Office AI uses the browser's native screen-capture API and LiveKit's simulcast encoding to keep your share crisp even on slower connections.
In a solo-AI room, your shared screen is also previewed to the assistant, so you can ask it about exactly what you're looking at.
Turn on live captions to read what's being said in real time. The overlay has three states — off, transcript, and translation — so a multilingual team can follow along in another language on the fly. Captions are driven by the server transcription pipeline; see Transcription for how transcripts are produced, saved, and made available to the AI.
Your Office AI does not have a one-tap "record this call" button. "Recording" means the server-side transcription pipeline, which produces text — captions and saved transcripts — rather than a video file. What's saved, and who can access it, is controlled by the transcription access policy.
If you're signed in on more than one device, you can beam things between them with remote commands — start a screen share, open a URL, or hand off an AI chat to another device. You can even drag a participant, skill, agent, knowledge folder, canvas, research report, workspace, or integration onto a device to send it there.
When you talk to the AI by voice, a full-screen overlay shows a waveform and live transcript and lets you barge in to interrupt. A persistent mini call module keeps the session handy as you move around the app.
If your camera or mic aren't working, check that your browser or OS has granted Your Office AI permission to use them. On macOS: System Settings → Privacy & Security → Camera / Microphone, then reload. For network issues, ensure your firewall isn't blocking the UDP ports LiveKit uses; it falls back to TCP/TURN for NAT traversal when a direct connection isn't possible.