The Voice Agent lets you interact with your connected apps, agents, and workspaces using natural voice commands. It’s built for speed—no clicks, just say what you need, and the Voice Agent gets it done across all your integrations.

​
What You Can Do

  • Connect and switch agents
    Pick from your available agents. The Voice Agent disconnects from the current session before switching—no duplicate sessions or confusion.
  • Ask or command anything
    Speak naturally. The Voice Agent interprets broad, even vague, requests. You don’t need to be exact; it figures out what you want using context.
  • Control the mic
    Mute, unmute, start, and stop conversations with a click or command. Status indicators show if you’re connected, connecting, or idle.
  • See agent details
    Once you pick an agent, you see its name, purpose, category, and unique motion (whatever that means in your setup).
  • Quick troubleshooting
    If something’s off (API issues, missing token), you’ll get a clear error and a retry button.
    Pro tip: Set up your environment variables: HUME_VOICE_API_KEY and HUME_SECRET.

​
Voice Agent Features

  • Status indicators:
    Instantly see if you’re connected, connecting, or waiting for an agent.
  • Conversation controls:
    Start/stop a session, mute/unmute the mic, and send text messages if you need to type instead.
  • Agent selector:
    Can’t decide? Easily swap agents, even mid-session.
  • Real-time feedback:
    Prompts and hints tell you what’s going on (“Speaking with X”, “Connecting…”, etc.)

​
Built-In Tools

When you talk to the Voice Agent, it can perform tasks using these built-in integrations (if you’ve connected them):
  • AI Search – Search the web, do research, find recent news, local results, and more.
  • Gmail – Read, send, or reply to emails; manage drafts, threads, contacts.
  • Google Drive/Docs/Sheets/Calendar/Meet – Anything you’d expect: read, edit, search, or manage files, events, meetings.
  • Slack – Message, search, or organize your workspace.
  • Notion, LinkedIn, YouTube, Reddit, Google Maps, Twitter, Microsoft Teams, Zoom, Canva, Calendly, Outlook, Confluence, Jira, HubSpot, Moz, Apollo, Klaviyo – All supported for direct actions via voice. If you connected it, just ask.
You’ll get a clear message if a tool is temporarily unavailable or not connected.

​
How to Use

  1. Open the Voice Agent module.
  2. Select an agent.
    • Agents are listed. If you switch, the previous session closes.
  3. Start a conversation.
    • Hit “Start” or use your voice to begin.
  4. Speak your command.
    • Example: “Show me my latest Gmail threads,” “Schedule a Zoom call for tomorrow,” “Search for the latest trends in AI accounting.”
  5. Check agent status and feedback.
    • Watch the status indicator or read the quick prompts.
  6. Stop or mute as needed.
    • Pause, mute, or end the session at any time.

​
Troubleshooting

  • Tool not responding?
    You’ll get a direct error. Try again or check your connection.

​
Best Practices

  • Don’t overthink your command. The Voice Agent is built to figure it out.
  • You can switch agents anytime, but current sessions will disconnect first.
  • If you get stuck or see an error, check your API keys or connected apps, then hit “Retry.”
You control everything from one place. Voice Agent is the fastest way to get things done across your apps—no typing, no tab-switching, just say what you want and let it work.