Productivity Analyser

Chrome extension + cloud dashboard for recruitment & staffing team productivity. The extension is the only agent the recruiter installs; everything syncs to a central Next.js cloud app where each recruiter sees their own day and admins see the whole team.

A native Windows companion agent is planned to cover the gap the extension cannot reach (non-browser app time).

What you get

Per-recruiter daily analytics (dashboard)

• Active time · Idle time · Total time
• LinkedIn time · Naukri time · Other browser time
• LinkedIn: profiles viewed (with name + title + dwell time + view count), connections sent, messages sent, searches run
• Naukri: candidate profiles viewed (name + title + dwell + views), CV downloads, contact reveals, searches
• Hourly timeline (stacked LinkedIn / Naukri / Other / Idle)
• Top domains visited

Team view (admin)

• All recruiters, same day, one table: active / idle / LinkedIn / Naukri / connections / messages / profiles / CVs
• Team-wide totals at the top

Repo layout

tray/          Go Windows tray agent (single static .exe, ~7 MB)
               foreground window + idle polling → /api/ingest/desktop
cloud/         Next.js 14 app (TypeScript, App Router)
  src/app/api/                 ingest + analytics endpoints
  src/app/dashboard|linkedin|naukri|domains|timeline|team/  pages
  src/lib/                     db (SQLite), auth (API key), analytics
extension/     Chrome + Firefox MV3 extension (shared source)
  src/
    background.js              per-URL dwell + idle + outbound batch flush
    content-linkedin.js        profile / connect / message / search events
    content-naukri.js          profile / CV / contact / search events
    options.html|js            configure cloud URL + API key
    popup.html|js              status + quick dashboard link
  manifests/
    chrome.json                MV3 w/ service_worker background
    firefox.json               MV3 w/ scripts background + gecko settings
  build.js                     copies src/ + correct manifest into dist/{chrome,firefox}

Architecture

┌──────────────────────────┐       HTTPS        ┌──────────────────────────┐
│  Chrome Extension (MV3)  │──X-PA-Key header──▶│   Cloud (Next.js)        │
│  - per-URL dwell         │                    │   /api/ingest/dwell      │
│  - linkedin content.js   │                    │   /api/ingest/event      │
│  - naukri content.js     │                    │        │                 │
│  - 30s batched flush     │                    │        ▼                 │
└──────────────────────────┘                    │   SQLite (swap to PG)    │
                                                │        ▲                 │
┌──────────────────────────┐                    │        │                 │
│  Recruiter / Admin       │◀──── Cookie ──────▶│   /dashboard /team ...   │
│  browser                 │                    │                          │
└──────────────────────────┘                    └──────────────────────────┘

Running the cloud app

cd cloud
npm install
# optional: set the admin key so you don't need to read server logs
export PA_ADMIN_KEY=supersecret-admin-key
npm run dev          # http://localhost:3000

On first run, an admin user is seeded and its API key is printed in the console (or set to PA_ADMIN_KEY if provided). Paste that key on / to sign in.

Creating recruiter users (admin only)

curl -X POST http://localhost:3000/api/users \
  -H "X-PA-Key: <admin-key>" \
  -H "Content-Type: application/json" \
  -d '{"name":"Priya Sharma","email":"priya@acme.com","team":"Tech Hiring"}'

Response includes a fresh api_key. Hand it to the recruiter; they paste it into the extension Settings page.

Installing the extension

The extension ships as a single source tree (extension/src/) with two manifests. A tiny build step produces loadable bundles for each browser:

cd extension
node build.js          # writes extension/dist/chrome/ and extension/dist/firefox/

Chrome / Edge / Brave / other Chromium

1. Open chrome://extensions → enable Developer mode
2. Load unpacked → select extension/dist/chrome/
3. Click the extension icon → Settings → paste cloud URL + personal API key → Save & test
4. Popup should now say Connected

Firefox

1. Open about:debugging#/runtime/this-firefox
2. Load Temporary Add-on… → select extension/dist/firefox/manifest.json
3. Click the extension icon → Settings → paste cloud URL + personal API key → Save & test
4. Popup should now say Connected

Temporary add-ons disappear on restart. For permanent install inside an org, sign the XPI via addons.mozilla.org (self-hosted distribution) or deploy it via an enterprise policy JSON.

Data starts flowing immediately. Background batches flush every ~30 s; short network outages are buffered (up to 500 records each for dwell + events). Any change to shared source in extension/src/ propagates to both Chrome and Firefox bundles on the next node build.js.

Deploying the cloud app

• Single VM (simplest): npm run build && npm start behind a reverse proxy. SQLite file lives under cloud/data/productivity.db. Persist that directory.
• Postgres later: replace cloud/src/lib/db.ts with a pg client; schema in that file maps 1:1.
• Vercel: requires swapping SQLite for a hosted DB (Neon, Supabase, Turso). The route handlers are already runtime = "nodejs".

What the extension captures vs. what still needs native

Capability	Extension	Windows agent
Per-URL / per-domain time	✅	—
LinkedIn profiles, connections, messages, searches	✅	—
Naukri profiles, CV downloads, contacts, searches	✅	—
Chrome idle (no activity in browser)	✅	—
Time in Outlook / Teams / Excel / Zoom / ATS desktop apps	❌	✅ (shipped)
Active window title outside Chrome	❌	✅ (shipped)
System idle when Chrome closed	partial (via alarms)	✅ (shipped)
Edge time	✅ (Chrome bundle works)	—
Firefox time	✅ (Firefox bundle)	—
Safari time	❌	✅
Screenshots / keystroke counts	❌	optional

Once we see how much of the real workflow sits inside Chrome, the Windows agent can be scoped precisely — it will just POST to the same /api/ingest/* endpoints.

Compliance

• Deploy only with recruiters' informed written consent.
• The extension is self-monitoring: it only records pages the user themselves navigates to. It does not crawl, enumerate, or scrape LinkedIn/Naukri beyond what the recruiter visits, staying aligned with those platforms' terms.
• All data is stored on your own server / DB. No third-party analytics.

Roadmap

1. Weekly / monthly rollups + per-recruiter targets
2. CSV / PDF daily export
3. Windows companion agent → shipped as tray/ (Go + Win32, see tray/README.md)
4. Anti-tamper check (flags gaps in data stream)