Interview Prep AI

Interview Prep AI is a career tool. A web app that simulates live job interviews by voice, using your CV, a job description, interview type and language to generate questions and give instant feedback. Built for repeated practice ahead of a real interview, with a stated policy that job description data is not stored on its servers.

Interview Prep AI is a web app built for one narrow job: rehearsing spoken job interviews before the real one. The premise is straightforward. Instead of reading sample answers silently or asking a friend to play interviewer, you talk through a live interview with an AI and get feedback on how you did. The product positions itself as a personal interview coach, and the workflow is structured around a single session type rather than a broad career toolkit. Setting up a session takes three inputs. You bring your CV, either by pasting the text or uploading the file so it can be converted to text and saved. You then fill in the role details, which means the job description, the kind of interview you are facing and the language you want to practise in. From there you run as many iterations as you like, speaking your answers and receiving feedback immediately after. Because the CV and job description shape the questions, the practice is anchored to a specific application rather than generic interview trivia. Language selection is a notable inclusion. Candidates interviewing in a second language can rehearse in that language, which puts the tool in the same conversation as general speaking-practice apps, though here the context is hiring rather than conversation fluency. On privacy, the site states that information such as the job description is not stored on its servers, which matters for anyone pasting details of an internal role or an offer in progress. Where it fits: this is a focused practice tool, not a resume builder or a job board. It competes with mock interview services that pair you with a human, and its advantage is repetition without scheduling. The trade-off is that feedback quality depends entirely on the model, so treat the output as a rehearsal aid and a way to catch filler and rambling, not as a verdict on whether you will get the job.