Listen

Purpose

Capture and transcribe what’s being played/spoken so the model can reason about the content (words, phrases).

Syntax

listen

Example:

1. playVoice: "apple"
2. listen
3. Check that the remembered text is exactly ‘apple’

Note:

• Use when you care about what was said (exact words, phrases, numbers, etc.).

• Follow‑up prompts should refer to the remembered text/content, not to “checking audio” (to avoid confusion with the ‎`check audio` tool).

• Support for this command is still on Beta. Please reach out if this is critical to your tests.