Automate Job Application

Enterprise-grade automated job application system (Bun + TypeScript, Express, Prisma, Redis Streams + Pub/Sub, Next.js).

Stack

• Runtime: bun
• API: express + ws
• Queue: redis streams (jobs) + redis pub/sub (events)
• Database: postgres + prisma
• Frontend: next.js (App Router)

Repo layout

• apps/api API server
• apps/worker Queue worker
• apps/web Dashboard UI
• packages/shared Shared types
• prisma Prisma schema and migrations

Prerequisites

1. bun installed
2. docker running

Quick start

1. Start dependencies

docker compose up -d

2. Install deps

bun install

3. Configure environment

cp .env.example .env

4. Generate Prisma client and migrate

bun run db:generate
bun run db:migrate

5. Install Playwright browsers (first time only)

bunx playwright install chromium

6. Run the stack

bun run dev

Docker (all services)

Build and run everything (API + worker + web + Postgres + Redis):

docker compose up --build

The API container will automatically run prisma migrate deploy on startup.

Stop all services:

docker compose down

Scripts

• bun run dev start API + worker + web
• bun run dev:api start API only
• bun run dev:worker start worker only
• bun run dev:web start web only
• bun run db:generate generate Prisma client
• bun run db:migrate run Prisma migrations

Environment

Single env file at repo root: .env

Notes

• WebSocket endpoint: ws://localhost:3001/ws?token=dev-token
• API endpoint: http://localhost:3001
• HEADLESS=false opens a visible browser for automation.
• BROWSER_CHANNEL=chrome uses your installed Chrome (optional).
• AUTO_SUBMIT=false stops at final review so you can submit manually.
• First run requires login in the opened browser (LinkedIn/Indeed).
• MANUAL_HOLD_MS controls how long the review window stays open.
• AI_SCORING=true enables Gemini scoring (limit with AI_SCORE_LIMIT).
• AI_ANSWER_ENABLED=true enables AI answers for application questions.
• AI_ANSWER_LIMIT=3 caps AI answers per form step.
• COVER_LETTER_ENABLED=true generates a cover letter per application.
• RESUME_TAILOR_ENABLED=true saves a tailored resume snapshot per application.
• AI_PROVIDER=auto chooses the LLM provider (openai, gemini, auto).
• OPENAI_API_KEY enables OpenAI usage (server-side only).
• OPENAI_MODEL sets the OpenAI model (default: gpt-4o-mini).

Discovery

• Discovery runs every 24h by default (DISCOVERY_INTERVAL_MS).
• Scheduled sources: Remotive + Arbeitnow (public APIs).
• Automation sources: LinkedIn + Indeed (via Playwright, requires login).
• Configure preferences in the dashboard: /dashboard/settings.
• API endpoints: POST /users, GET /users?email=..., POST /preferences, GET /preferences/:userId, POST /automation/start.

Automation

• Start applying triggers browser automation for LinkedIn + Indeed.
• AUTO_SUBMIT=false stops at final review so you can submit manually.
• First run requires login in the opened browser; session is reused.

Resume parsing

1. Add GEMINI_API_KEY to .env.
2. (Optional) Add OPENAI_API_KEY and set AI_PROVIDER=openai or choose OpenAI in the dashboard.
3. Open /dashboard/settings, upload a resume PDF, click Upload & parse.

AI scoring & cover letters

• The worker derives keywords/roles from your parsed resume if preferences are empty.
• AI scoring uses Gemini (if enabled) for up to AI_SCORE_LIMIT jobs per run.
• Cover letters are generated per application and saved in the database.
• Tailored resumes are saved per application in artifacts/resumes/.
• AI answers autofill short text, dropdown, and radio/checkbox questions.

Documents

• Download tailored resumes and cover letters as PDF from /dashboard/resume-viewer.
• Generate interview prep per application and view it in the same panel.

Quick setup (example):

curl -X POST http://localhost:3001/users -H 'Content-Type: application/json' \\
  -d '{"email":"you@example.com","fullName":"Your Name"}'

curl -X POST http://localhost:3001/preferences -H 'Content-Type: application/json' \\
  -d '{"userId":"<USER_ID>","roles":["software engineer"],"keywords":["react","typescript"],"locations":["remote"],"remote":true}'

Progress

• Monorepo scaffold with Bun workspaces
• Express API with WebSocket event streaming
• Redis Streams job queue + Redis Pub/Sub for events
• Prisma schema (core + resume/cover letter/import/notifications)
• Worker consumer loop for stream jobs
• Discovery scheduler (daily) with Remotive + Arbeitnow sources
• Next.js dashboard shell (Overview, Jobs, Queue, Logs, Settings)
• Discovery UI (create user + preferences)
• Discovery results UI (matches + jobs + runs)
• Start applying button (automation trigger)
• Playwright automation (LinkedIn + Indeed, stops at final review)
• Resume upload + Gemini parsing (profile extraction)
• AI scoring + cover letter generation
• Resume & cover letter viewer UI
• PDF export for tailored resume and cover letter
• Interview prep generation (AI)
• Auto-fill common form fields from resume profile
• Tailored resume snapshots per application
• AI answers for text + dropdown/radio/checkbox questions
• Live overview metrics + WebSocket log stream
• Queue status view backed by live application data
• Bulk job import endpoint + UI (paste multiple URLs)
• Email notifications on status updates (SMTP)
• Analytics dashboard (conversion + platform breakdown)
• Chrome extension (one-click save/apply)

Roadmap

• More platform adapters (Lever, Greenhouse, Wellfound)
• Resume PDF rendering (LaTeX or HTML/PDF)

Chrome extension

Location: apps/extension

Load it in Chrome:

1. Open chrome://extensions
2. Enable Developer mode
3. Click Load unpacked
4. Select apps/extension

Configure the API URL + User ID (Options page), then click Save & apply to queue a job.

Analytics

Dashboard: /dashboard/analytics

API: GET /analytics/summary?userId=...