Skip to main content

Search

Search problems, hooks, and pages

Back to About

Mock Interview

built with: Dexie · Web Speech API

The question bank is the app's own content, and a spoken answer and a written reflection are kept as two separate records per question.

Questions drawn from content that already exists

generateQuestions() pulls from the same notes, flashcards, and machine-coding problems every other feature already has, so the pool is always in sync with whatever content exists — there is no separate interview-question dataset to keep updated by hand.

It shares its live speech capture with Speak Up (the same use-speech-input hook, built on the typed SpeechRecognitionLike wrapper), applied to a structured chat sequence instead of open rehearsal.

No grading, on purpose

The UI says so directly: this is a reflection tool, not a scored assessment. A recap phase reviews the full transcript after the session, but nothing here produces a score.

Each answered question keeps its live spoken transcript and an optional written reflection as two distinct fields, so editing your thinking afterward never overwrites what you actually said in the moment.

Same Dexie pattern as everywhere else

mockInterviews is one more Dexie table behind the same repository + useLiveQuery pattern as Job Tracker and Speak Up, so a mock interview persists across reloads and stays in sync across open tabs with no extra code.

Important

Gotcha

The live spoken transcript and the post-hoc reflection are saved through different paths: the transcript is captured silently during the interview, while a later manual edit to the reflection toasts a confirmation — so the two writes look and feel different on purpose, matching how differently they happen.

Where it lives in the repo

  • src/components/mock-interview/mock-interview-view.tsx
  • src/components/mock-interview/mock-interview-chat.tsx
  • src/components/mock-interview/chat/use-speech-input.ts
  • src/data/mock-interview-pool.ts
  • src/db/mock-interviews.ts