All guides
Talking to Cody or Cloe by voice

Talking to Cody or Cloe by voice

~3 min read

Closr ships with a full voice conversation mode powered by ElevenLabs. You talk, your AI partner talks back. No typing. The same Cody (or Cloe, your call during onboarding) you use in chat is available as a voice partner. It's the fastest way to use Closr when your hands aren't free.

Where the entry point lives

In the chat input bar there's an audio waveform icon next to where you'd hit send. Tap it. That starts a voice session with your AI partner.

Today the icon isn't labeled. We know it's unobvious. That's on the rough-edges list from the alpha invite. Once you've used it once you'll find it again.

When it's actually useful

Voice mode is built for the moments when you can't sit at a keyboard:

  • •**Driving between showings.** "What does my afternoon look like? Do I have anything from Sarah on 142 Eastern? Draft a quick text to my 4pm telling them I'm running 10 minutes late."
  • •**Hands-free briefings before a showing.** "Brief me on 88 Wilson. Asking, recent activity, anything I should know about the seller." You get a spoken summary of the deal context, recent comms, and any flagged risks.
  • •**Walking the block after a listing appointment.** "Log that I just met with the Smiths. They're motivated, want to list around $1.49M, willing to do staging. Add a follow-up for tomorrow morning."
  • •**End of day, on the couch.** "What slipped today? What needs my attention tomorrow?"

The same tools Cody has in chat are available by voice: looking up contacts, drafting emails or texts, running a CMA off comps you describe, advancing deals, logging activities. Anything you can ask in chat, you can ask out loud.

How a session works

You speak. ElevenLabs streams the audio to Closr's voice agent. The agent picks the right tool (or just talks back), executes, and replies in your AI partner's voice. Tools that produce visible output (like a draft email or a deal card) still get rendered in the chat UI for you to see when you come back. So you can run a CMA by voice while driving, then read the comp table on the phone when you stop.

End the session by tapping the icon again or saying "that's it" / "we're done".

What's rough today

  • •**The entry point is unlabeled.** Noted. Inline label is queued.
  • •**No wake word.** You start sessions manually. We're not always-listening, on purpose.
  • •**Voice cloning is separate from voice mode.** Voice mode uses one of two default ElevenLabs voices (Cody or Cloe). Cloning your own voice for outbound calls is set up in Settings → Profile and applies to the AI voice agent that answers your business line, not to the back-and-forth voice partner described here.

Where to go next

  • •`voice-profile`: training Cody on your writing voice (separate from voice mode)
  • •`sms-and-calls`: the AI voice agent that answers your business line
  • •`getting-started`: the rest of the chat surface