Capture spoken answer

Plays a prompt, transcribes the caller’s free-form spoken answer into a variable, then continues. Use it to collect open-ended input (name, reason for cal…

What it does

Plays a prompt, transcribes the caller's free-form spoken answer into a variable, and continues. Use it to collect open-ended input — a name, the reason for calling, feedback — for a later node or an integration.

The transcript is stored under Store transcript in (default recognizedText) and is always also available as vars.recognizedText. Reference it downstream as {{vars.<name>}} — e.g. pass it to an apiCall node or speak it back with say.

Unlike gatherSpeech, this node doesn't branch on the answer — it has a single onward edge (call.speech.ended).

When to use it

• "What's your reason for calling?" → store → send to your CRM via apiCall
• Capture a spelled name or order number for an agent screen-pop
• Open-ended feedback you want as text, not an audio recording

Notes

• Requires a recognizer configured under Admin → Speech recognition.
• STT usage is metered per audio minute against the recognizer's rate.