InterviewForge

InterviewForge is a local-first Python tool that turns a recording of a job interview into a structured PDF review report. The problem it solves is the difficulty of learning from past interviews: most people just move on without a reliable record of what was asked and how they actually answered. With InterviewForge, you point it at a local video or audio file from an interview, it extracts the audio, transcribes it using a local speech recognition tool, then uses an AI agent to clean up the transcript, identify the interviewer's questions, evaluate each answer with a quality label (strong, passable, risky, or weak) and a score out of five, and finally compiles everything into a formatted PDF using LaTeX. The report includes a summary page, detailed question-and-answer cards with timestamps, a section on high-risk technical points, suggested better answers with cited sources, and follow-up study materials. Everything stays on your machine by default. No audio, transcript, or report is uploaded anywhere. It also works as a skill that AI coding assistants like Codex or Gemini CLI can install and run on your behalf.