On face-detected
The On face-detected handler fires when the kiosk camera detects a human face approaching. Use it to greet users before they speak.
The On face-detected handler fires when the kiosk camera detects a human face within range. It is only triggered if the camera has been enabled in the kiosk configuration – see Enabling Camera Detection below. This allows your virtual being to proactively greet approaching users before they say a word – creating a more natural, welcoming experience at the kiosk.
When It Runs
This handler is triggered by the predefined face-detected signal, which fires when:
- The camera is enabled in the kiosk's on-device settings (see Enabling Camera Detection)
- The camera detects a human face in the frame
- A new conversation is not already in progress (to avoid re-greeting a user mid-conversation)
This handler is Kiosk only. It requires the camera to be enabled in the kiosk app's on-device settings – this is not a Dashboard profile setting. It does nothing on web widget deployments.
Common Uses
- Proactive greeting – Activate a greeting flow the moment a user approaches, without waiting for them to speak
- Wake from idle – Transition the kiosk from its idle animation/state to an active ready state
- Start session – Combine with
On conversation-startto create a seamless approach-to-conversation experience
Typical Node Pattern
Signal Handler (On face-detected) ──► Activate Flow (Greet)Or with an idle state check:
Signal Handler (On face-detected) ──► Is Flow Active? (Greet) ──► [No] ──► Activate Flow (Greet)
└► [Yes] ──► Abandon ProcessingThe second pattern prevents re-triggering the greeting if it's already playing.
Available Nodes
| Node | Category | Purpose |
|---|---|---|
| Continue Processing | Exit | Pass to Conversational Flows |
| Abandon Processing | Exit | Drop the event |
| Activate Flow | Exit | Trigger a specific Conversational Flow |
| Check Variable | Context | Read a variable |
| Store Variable | Context | Write a variable |
| Route by Profile | Logic | Branch by touchpoint |
| Route by Signal | Logic | Branch by signal |
| Is Flow Active | Logic | Check if a flow is already running |
| Trigger Automation | Integration | Fire an Automation |
| Comment | Other | Canvas annotation |
Example: Proactive Greeting with Guard
Prevent the greeting from firing if the kiosk is already in an active conversation:
- Signal Handler (On face-detected)
- Is Flow Active (check
MainorAgentflow)- Active → Abandon Processing (user is already in conversation, do nothing)
- Not active → Activate Flow (Greet)
Enabling Camera Detection
Camera-based face detection must be enabled in the kiosk settings before this handler will fire:
- Open kiosk application
- Go to app settings (tap three times in the bottom right corner of the screen)
- *Enter PIN if profile is protected by PIN
- In wake mode select: camera
- Save and close settings
Without this setting, the face-detected signal will never fire and this handler will remain dormant.
Hardware Requirements
For reliable face detection, follow the camera recommendations in the MetaHuman Kiosk hardware setup guide. Camera placement, lighting conditions, and frame rate all affect detection reliability.