Interview Copilot (Local, Unpublished)

Interview Copilot listens during a live call and streams AI-generated, C-level answer suggestions in real time — grounded in your own resume and the job description, not generic advice. HOW IT WORKS - Listens through your microphone during a call (works with speaker or laptop audio — not compatible with headphones, since the mic needs to physically hear both sides of the conversation). - Automatically detects when a question has been asked (a pause in speech triggers it — no button needed for that part). - Streams a structured, step-by-step suggested answer into a small panel on the page, written in your voice using your real background. - After each answer, the screen freezes on purpose so your own voice reading it back isn't mistaken for a new question — press any key to clear it and move on to the next one. SETUP (2 minutes, one time) 1. Get a free Gemini API key at https://aistudio.google.com/apikey (Google's own AI Studio — no cost to create a key). 2. Click the extension icon once. Since no key is saved yet, this opens the Settings page automatically. 3. Paste your Gemini API key, then your resume, the job description, and the role/company you're interviewing for. Click Save. 4. During your interview, click the extension icon again on the tab you're using. Allow microphone access if prompted (only needed once). A panel appears — it will start listening automatically. YOUR OWN API KEY, YOUR OWN DATA This extension does not have a backend server. Your API key and the interview context you enter are stored only in this browser's local extension storage, never sent to us or synced anywhere. When you ask a question, your resume/JD/question text is sent directly from your browser to Google's Gemini API to generate the answer — the same request your own key would make if you called Gemini's API yourself. WHO THIS IS FOR This is a practical prep tool built for one specific situation: an IT/technology leadership interview, where you want to speak with the kind of specificity, numbers, and executive framing that these interviews reward. It works best when you fill in real, specific details in your resume and the JD, since the AI is instructed to ground every answer in what you actually gave it rather than invent generic filler. LIMITATIONS, HONESTLY - Requires a Gemini API key you provide and pay for yourself (Google's free tier is generous, but usage beyond it is billed to your key). - Microphone-based only in this version — it cannot distinguish your voice from the other speaker's with certainty, since both come through one microphone. The "freeze until keypress" behavior exists specifically to manage this. - English speech recognition only.