Interview setup

The setup is what turns generic AI output into answers that sound like you and reference work you actually did. It starts with Live copilot or Mock interview.

The setup form — profile through role — then start.

What to fill in

Live copilot or mock interview

The first choice on the form. Live copilot stays with you in the real interview — we listen and draft answers as you go. Mock interview flips the roles: the app asks the questions, listens to you, then scores the round. Mock is Pro and higher; Free shows an unlock plaque that opens Settings → Plan.

Mock is verbal-only. Verbal is selected for you, and you pick 20 or 30 minutes. Live copilot still lets you pick Verbal, Coding, System design, Project, or any combination. Mock interviews →

Role

The position you are interviewing for. This shapes vocabulary and the kind of examples the assistant reaches for. Be specific — "Senior Backend Engineer" produces better output than "Engineer".

Job description

Paste the posting. The assistant mirrors the language the company uses, which matters more than people expect: interviewers tend to score answers higher when they hear their own framing reflected back.

Resume and personal profile

You do not have to fill the profile in by hand. Upload a resume (.pdf, .doc, .docx, .txt, .md, or .rtf) or paste it in and choose Extract with AI — the app fills a stack-agnostic summary of your role, domain, impact, and seniority, which you can then edit.

Named tools and languages stay off the profile on purpose. Attach the full resume under Add more → Resume when you want the assistant to cite specific stacks.

Optional add-ons

Resume, Q&A, and additional context sit under Add more. Resume takes a file or pasted text. Q&A and additional context take notes, or a snippet from a past interview. They are extra grounding, not required.

Support types

Choose which interview types this session needs — Verbal, Coding, System design, Project, or any combination. Only the ones you pick are prepared, so leaving out types you will not use keeps the workspace uncluttered. Mock interviews skip this picker because they always run as verbal. Project cannot be added after the session starts.

Project folder

Required when Project is selected. Pick a local folder to bind it to the session. The app skips junk while checking the folder, and Start stays blocked until it finds at least one readable source file. During the round, the Project surface reads the live tree with tools.

Programming language

Required when Coding is selected. Solutions come back idiomatic to the language you pick rather than translated out of Python.

CC#C++DartElixirGoJavaJavaScriptKotlinPHPPythonRRubyRustScalaSQLSwiftTypeScript
Models are not part of the setup form
Each surface uses the model configured in Settings → General → AI models, applied when the interview starts. Set your per-surface models there before the round. See Choosing models for a recommended pairing.

Saving and reusing setups

Start an interview and the setup saves itself — there is no save button to remember. Each unique configuration appears once under Recent interview setup in the left sidebar; click the row and the form fills back in exactly as you had it. For a company loop with four rounds over two weeks, that is one row you click four times — while your interview history keeps every session separate so you can look back at what was asked.

Click a recent setup and the form fills back in.
Setup titles come from the role
A saved setup is titled with its role's company and title — "Stripe — Senior Backend Engineer". Make the role record specific when you create it: "Backend Engineer — Technical Screen" beats "Engineer" when you are three loops deep across two companies.
User Guide