Case CompendiumX

This is a Next.js project bootstrapped with create-next-app.

Getting Started

First, run the development server:

npm run dev
# or
yarn dev
# or
pnpm dev
# or
bun dev

Open http://localhost:3000 with your browser to see the result.

Environment Variables

Create a .env.local file with:

NEXT_PUBLIC_FIREBASE_API_KEY=...
NEXT_PUBLIC_FIREBASE_AUTH_DOMAIN=...
NEXT_PUBLIC_FIREBASE_PROJECT_ID=...
NEXT_PUBLIC_FIREBASE_STORAGE_BUCKET=...
NEXT_PUBLIC_FIREBASE_MESSAGING_SENDER_ID=...
NEXT_PUBLIC_FIREBASE_APP_ID=...
GEMINI_API_KEY=...
# Optional override (default: gemini-2.5-flash-lite)
GEMINI_TRANSCRIBE_MODEL=gemini-2.5-flash-lite

GEMINI_API_KEY is used only in the server route (/api/transcribe) for audio-to-text.

You can start editing the page by modifying app/page.tsx. The page auto-updates as you edit the file.

This project uses next/font to automatically optimize and load Geist, a new font family for Vercel.

Backend (Firebase)

The Firebase project (compendium-x) is version-controlled from this repo:

File	Purpose
firebase.json	Manifest — tells the Firebase CLI what this repo deploys (rules, indexes, emulators).
.firebaserc	Pins this repo to the compendium-x Firebase project.
firestore.rules	Security rules controlling who can read/write each Firestore collection.
storage.rules	Security rules for uploaded files (recordings, workspace images).
firestore.indexes.json	Composite indexes Firestore needs to answer multi-field queries.
functions/	Firebase Cloud Functions — currently hosts the long-running transcription job.
serviceAccountKey.json	Admin credentials for server-side scripts. Never commit. Gitignored.

One-time setup

1. Install Java (needed by the Firestore + Storage emulators):

   brew install --cask temurin
   

2. Log in to Firebase so you can deploy:

   npx firebase login
   

3. Configure Vercel env vars so server routes can authenticate as Firebase Admin:
   • Vercel dashboard → compendium-x project → Settings → Environment Variables
   • Add FIREBASE_SERVICE_ACCOUNT_JSON — paste the entire contents of serviceAccountKey.json as a single line, scoped to Production + Preview
   • Trigger a redeploy so it takes effect
   • Locally this variable is not needed — the file at the repo root is used automatically.
4. Upgrade Firebase to Blaze (pay-as-you-go) — required to deploy Cloud Functions:
   • Firebase Console → Project Settings → Usage and billing → Modify plan → Blaze
   • Set a $5/month budget alert at the same time (Firebase Console → Usage and billing → Details and settings → Budget)
5. Store the Gemini API key as a Firebase secret (used by the transcription Cloud Function):

   npx firebase functions:secrets:set GEMINI_API_KEY
   # paste the same value you have in .env.local, hit enter
   

6. Deploy the Cloud Function (one-time + after function changes):

   cd functions && npm install && cd ..
   npx firebase deploy --only functions
   

Local development with emulators (recommended)

Instead of npm run dev (which writes to production Firestore), use:

# Terminal 1 — start the local Firebase emulators
npm run emulators

# Terminal 2 — start Next.js pointed at the emulators
npm run dev:emulators

The Emulator UI runs at http://127.0.0.1:4000 — you can inspect Firestore docs, Storage files, and emulated Auth users live.

Deploying rules / indexes

npm run rules:deploy       # deploys firestore.rules + storage.rules
npm run indexes:deploy     # deploys firestore.indexes.json

Deploying the transcription Cloud Function

cd functions
npm install
npm run build
cd ..
npx firebase deploy --only functions

Function logs (last hour) when something's misbehaving in prod:

npx firebase functions:log

How transcription works

1. Candidate finishes recording → audio uploads to Firebase Storage.
2. Client calls POST /api/sessions/[lobbyId]/recording with metadata. This writes recording.transcriptStatus: 'pending' to the session doc.
3. Firestore-triggered Cloud Function transcribeRecording fires, downloads the audio, hands it to Gemini, and writes the transcript back to the session doc when done.
4. Dashboard / evaluation pages display the transcript via existing real-time listeners — no client polling needed.
5. If transcription fails, the candidate's workspace shows a Retry Transcript button that calls POST /api/sessions/[lobbyId]/recording/retry-transcript to re-queue.

The function has a 9-minute timeout (plenty for 45-minute audio at Gemini Flash speeds), runs on Blaze pricing, and is regional us-central1. Logs live in Firebase Console → Functions → Logs.

To inspect what's currently deployed in production (uses serviceAccountKey.json):

npm run rules:dump

Learn More

To learn more about Next.js, take a look at the following resources:

• Next.js Documentation - learn about Next.js features and API.
• Learn Next.js - an interactive Next.js tutorial.

You can check out the Next.js GitHub repository - your feedback and contributions are welcome!

Deploy on Vercel

The easiest way to deploy your Next.js app is to use the Vercel Platform from the creators of Next.js.

Check out our Next.js deployment documentation for more details.