Verbal copilot

On the call, the verbal surface captions the question as it is asked, then drafts an answer from your resume. You speak it yourself. It covers behavioral questions and any other conversation rather than a problem.

This is the verbal surface of the interview copilot.

The verbal surface mid-answer: the transcribed question at the top and a streaming STAR-formatted response below it. — A behavioral answer streaming in.

Using it

Press the shortcut for Verbal’s slot in this interview (often ⌘⌥1 when it is first) to focus that surface. The interviewer’s words land in the draft as they are transcribed. Type or paste when you want a different question. Send posts whichever is there.

First words of a suggested answer: under 100 ms for Verbal; up to about 5 seconds for Live coding, System design and Project.

Grounding

With a resume and job description attached in the setup, answers reference your real experience. Without them you get competent but generic output — which is exactly what an interviewer has heard forty times that month.

Per-thread instructions

Each surface is its own thread, and you can give a thread standing instructions from the control bar — "keep answers under a minute", "assume a staff-level audience", "do not mention my current employer by name". These apply to every response on that thread and sit on top of your answer style, without changing the prompt saved on that interview’s custom scenario.

Read it, do not recite it

The strongest use of this surface is to glance at the structure and then say it in your own words. Reading verbatim is audible — the cadence gives it away.