Skip to content

Receptionists & Kiosks

This screen shows how your public Kiosk receptionist is being used by visitors: how many sessions ran, when, over which channels, and who’s configured to greet people. The counts come from real kiosk sessions on this machine, so viewing them costs no tokens. When you’re ready to share, one click builds a branded activity report you can copy or download.

How to get here

Open Reports → Receptionists & Kiosks (or wherever it’s surfaced in your sidebar). To configure the kiosk itself — pick the greeter, voice, and language — go to Settings → 🖥️ Kiosk.

What you’ll see

  • An entitlement banner confirming whether the Kiosk add-on is active and how many live kiosks are licensed.
  • Metric tiles — total visitor sessions, today, last 7 days, average per day, licensed kiosks, and when the last visitor came by.
  • By channel — once telemetry exists, a breakdown of voice vs. text and in-app vs. remote sessions, plus the top languages seen.
  • Front desk inbox — any messages, bookings, or check-ins the receptionist captured (these also land in that worker’s Inbox).
  • Configured receptionist — the worker on duty, their voice engine, and default language.
  • Recent visitor sessions — the latest sessions with timestamps.

The shareable report

Click 📄 Generate report to open a clean, Markdown activity summary covering the licence, the configured receptionist, visitor usage, and channels. From there you can Copy it or ⤓ Download .md to hand to a stakeholder.

In practice

Try the in-app kiosk before buying the add-on. It works without the entitlement, so you can configure a greeter, talk to it as a visitor would, and decide whether the experience is right for your front desk — then license live kiosks once you know.

The metrics answer questions you can’t answer from memory. Two that come up:

  • Is anyone actually using it? Sessions today versus the 7-day average, and when the last visitor came by. A kiosk in a quiet corner of an office produces a very different number than one at the entrance, and that’s worth knowing before you renew.
  • Are they speaking or typing? The channel breakdown tells you whether the voice setup is earning its cost or whether visitors default to the keyboard.

Generate report is for the person who’s paying. It produces a clean Markdown summary — licence, configured receptionist, usage, channels — that you can hand to a client or a manager without screenshotting the screen. Copy or ⤓ Download .md.

Captured messages land in two places. Bookings and check-ins the receptionist takes show in the front-desk panel here and in that worker’s Inbox, so triaging them is normal inbox work rather than something that only exists on this screen.

Empty numbers mean it hasn’t run, not that tracking is broken. Configure and launch under Settings → 🖥️ Kiosk, hold one session as a visitor, and it will appear here — viewing costs no tokens, so check as often as you like.

🎓 Learn it hands-on: Kiosks: a public-facing agent

Terms in this page

  • Kiosk — a public-facing askTheodor front end a visitor can walk up to and talk with (in person or remotely).
  • Receptionist — the worker configured to greet visitors at the kiosk.
  • Visitor session — one visitor’s conversation with the kiosk receptionist.
  • Kiosk add-on — the paid entitlement (€50/mo per live kiosk) that licenses kiosks for production use.
  • Live kiosk — a kiosk running for real visitors, counted against your licensed cap.
  • Licensed cap — the number of kiosks your plan permits; a billing limit, not a performance one.
  • Entitlement — what your current licence allows (here, whether the kiosk add-on is on and for how many kiosks).
  • Telemetry — automatically recorded usage data, such as a session’s channel and language.
  • Channel (voice / text, in-app / remote) — how a visitor reached the kiosk: by speaking or typing, on this machine or over the network.
  • Front desk inbox — messages, bookings, or check-ins the receptionist captured during sessions.
  • Voice engine — the speech system the receptionist uses (e.g. a real-time voice model, or plain system voice/text).
  • Token — the unit AI models bill per; viewing these metrics spends none.
  • Markdown — lightweight text formatting used for the copied/downloaded report.