palaiologos1453/openinterview

OpenInterview is a locally-run AI mock interview tool aimed at people preparing for Java backend and AI application development job interviews or internships. Everything runs on your own computer: the question bank, the scoring, and the session history are all stored locally in a SQLite database. You only need an external LLM API key if you want the tool to generate a written summary report after a session, and even then the key stays on your machine. Setup on Windows uses a single PowerShell script that finds available ports, creates a Python virtual environment, installs dependencies, and starts both the FastAPI backend and a static HTML frontend. Once running, you open the app in a browser and select an interview mode and interviewer style before starting. The tool offers four interview modes: a comprehensive round that covers self-introduction, project walkthrough, computer science fundamentals, and system design, a pure Q&A round focused on drilling core knowledge, a resume-grilling round that probes the projects listed on your resume for authenticity and depth, and a system design round focused on requirements, capacity estimation, and tradeoff discussion. Each mode pairs with an interviewer style that controls the follow-up pattern. After a session, the tool produces a report covering practice questions, recommended answer structures, example answers, and a flashcard for each question showing score breakdown, missing points, and suggested rewrites. The question bank currently has over 700 Java backend questions and over 250 AI application development questions, drawn and reformatted from the JavaGuide project. Speech input and output default to the browser's built-in capabilities. For fully offline speech, local SenseVoice (transcription) and CosyVoice (text-to-speech) models can be configured separately.