Structure

Layers are inferred from where files sit in the tree, not from reading the code. A project that names its directories unconventionally will read oddly here — open the file browser to check anything the diagram implies.

Languages

Share of indexed source by file size. Binary and vendored files are excluded.

Dependencies

Declared in the repository’s manifests at the indexed commit. A declared package is not proof it is used, and runtime dependencies are listed first.

Feature verification

• Confidence meter / score displayed to userVerified

  personalized feedback... to assess and improve confidence

  Claimed on readmehigh confidence

  • components/Chat.tsx:209 — A Gauge component renders a live 'Confidence Meter' computed from the prosody-derived conf value
• Filler word detectionVerified

  tools to build confidence and improve interview performance (filler word tracking)

  Claimed on readmehigh confidence

  • components/Chat.tsx:98 — Splits transcript text and counts occurrences of a filler word set (um, uh, like, etc.), displayed as '# of filler words'
• Google OAuth sign-inVerified

  Google OAuth listed as a required credential / sign-in mechanism

  Claimed on readmehigh confidence

  • pages/api/auth/[...nextauth].js:28 — NextAuth is configured with GoogleProvider using GOOGLE_CLIENT_ID/SECRET
  • app/auth/page.tsx:78 — Sign-in page calls signIn('google') on button click
• Per-interaction data persisted to Firebase FirestoreVerified

  Firebase-backed storage of interview session data supporting feedback/history

  Claimed on Devposthigh confidence

  • components/Messages.tsx:45 — addDoc(collection(db,'questions'), data) persists each message with email, text, interview number, and prosody analysis scores
• Real-time AI mock interview via Hume voice/emotion APIVerified

  AI-powered real-time interview assistant that conducts mock interviews

  Claimed on readmehigh confidence

  • app/interview/page.tsx:8 — fetchAccessToken from hume package is called with HUME_API_KEY/HUME_SECRET_KEY and passed to the client component
  • components/Chat.tsx:174 — VoiceProvider from @humeai/voice-react is used with the access token and configId, wired to Messages/Controls/StartCall for a live voice interview
• Sentiment/tone-of-voice analysis during interviewVerified

  Athena analyzes sentiments and tone of voice during the interview to assess and improve confidence

  Claimed on readmehigh confidence

  • components/Chat.tsx:76 — gotData reads newMessageObject.models.prosody.scores from Hume's prosody model and derives a confidence value and top emotions
  • components/Messages.tsx:87 — msg.models.prosody?.scores is captured per message and stored as 'values' for each interaction
• Personalized post-interview performance dashboard with analytics and feedbackCode-supported

  comprehensive support ... personalized feedback, helping users build confidence, ensuring a well-rounded preparation experience

  Claimed on readmemedium confidence

  • app/performance/page.tsx:26 — Fetches real Firestore data via getSessionInfo, but the query hardcodes email 'saarnav@berkeley.edu' and interview 1 instead of using the signed-in user
  • app/performance/page.tsx:73 — chartData is overwritten immediately after fetch with hardcoded placeholder arrays (e.g. [1,3,5,7,9,11]), so the line chart does not actually reflect fetched data
  • app/performance/page.tsx:136 — 'Overall Feedback' text is a static hardcoded string, not generated from the interview data
• Resume upload / analysis featureCode-supported

  resume support as part of interview preparation tooling (app/resume page)

  Claimed on readmelow confidence

  • app/resume/page.tsx:14 — Drag-and-drop file uploader calls a Scrape() function on drop, but only logs file metadata; no visible parsing/analysis of resume content is wired to feedback
• Multiple named interview projects / history per user (dashboard of past interviews)Claimed only

  students build confidence in their job search via multiple mock interview sessions

  Claimed on readmemedium confidence

An AI agent derived these features from the project’s Devpost page and readme, then searched the code for each one. Verified features are backed by cited code; claimed-only features had no supporting code, which is not by itself proof a feature is missing.