Automated job application tool — find, score, and apply to LinkedIn positions hands-free.

---

How it works

1. Fill in your profile — target job titles, skills, resume, and location preferences
2. Click Search Jobs — the scraper finds matching LinkedIn postings and scores each one
3. Review results in the dashboard — each job is scored 0–100 against your resume by an LLM
4. Select jobs and click Apply to Selected — an AI agent fills out and submits each application, generating a personalized cover letter and PDF resume on the fly
5. Optionally configure a schedule to run searches automatically on a daily or weekly cron

Job tracking

Every job found by a search is saved and stays in your dashboard whether or not you ever apply to it — so the app works as a standalone job board even if you skip Easy Apply entirely.

• Status pipeline — each job starts at pending_review and can be moved to applied, rejected, or skipped manually from the dashboard (applying/failed are set automatically when the AI agent runs)
• Search + filter + sort — filter by status or workplace type (on-site/remote/hybrid), search by title/company/location, sort by score or recency
• Score at a glance — every job carries its 0–100 LLM match score, so you can triage without re-reading each posting
• Company history — the detail view shows how many times you've applied to or been rejected by that company before, and which titles
• One-click back to source — every job links back to the original LinkedIn posting
• Smart deduplication — skips jobs already in your dashboard by URL, and (if enabled) by matching company + title + location, so re-running a search doesn't flood you with the same postings; exclude keywords and companies you never want to see

Stack

Layer	Tech
Frontend + API	Next.js 16 App Router
Auth	Better Auth (email + password)
Job queue	BullMQ + Redis
Database	PostgreSQL + Drizzle ORM
Scraper	Playwright (LinkedIn)
AI Agent	LLM agent + Playwright MCP
Observability	Langfuse (self-hosted)

Monorepo structure

apps/
  web/        Next.js frontend + API routes (port 3000)
  worker/     BullMQ worker — runs scraper and AI agent
packages/
  api/        tRPC routers, services, BullMQ queue definitions
  db/         Drizzle schema + migrations + repository query functions
  automation/ LinkedIn scraper
  ai/         AI apply agent + resume PDF generator + LLM job scorer
  shared/     Shared utilities and constants (used by api + worker)

Getting started

Prerequisites: Docker with Compose v2.20+ (ships with Docker Desktop 4.22+)

1. Clone

git clone <repo-url>
cd applied

2. Configure

Get an API key from v0.dev/gateway (requires a Vercel account):

echo "AI_GATEWAY_API_KEY=your-key" > .env

Before exposing to the internet: also set BETTER_AUTH_SECRET (openssl rand -base64 32) and ENCRYPTION_KEY (openssl rand -hex 32) in .env.

3. Start

docker compose up -d

Migrations run automatically before the app starts. Everything else is pre-configured.

4. Open

Go to http://localhost:3000 and create an account.

Contributing

See CONTRIBUTING.md for local development setup, commands, and architecture notes.

Disclaimer

This tool automates interactions with LinkedIn in ways that may violate their User Agreement. Use it at your own risk. The authors are not responsible for any consequences including account suspension or legal action.