Self-hosted job search automation — scrape any career portal or use job aggregator, AI scoring against your profile, resume tailoring with custom themes, Telegram notifications and tracking in one system.
┌─────────────────────────────────────────────────────────────────────────────┐
│ JOB DISCOVERY │
│ │
│ Career Pages │ Aggregators │ Chrome Extension │
│ │ │ │
│ Any site via │ JobSpy: LinkedIn, │ Passive LinkedIn │
│ Playwright │ Indeed, ZipRecruiter, │ capture while │
│ │ Google Jobs │ browsing │
│ 11 ATS endpoints: │ │ │
│ Workday, Greenhouse │ LinkedIn Personal │ Save any job from │
│ Lever, Ashby, │ collections │ any page │
│ Oracle, Phenom, │ │ │
│ TalentBrew, Rippling│ Levels.fyi │ │
│ SmartRecruiters, │ │ │
│ + custom │ │ │
│ │ │ │
└───────────────────────┴────────────┬────────────┴───────────────────────────┘
│
▼
┌─────────────────────────────────────────────────────────────────────────────┐
│ PROCESSING │
│ │
│ Dedup ────── URL-hash dedup, tracking params stripped │
│ Filters ──── Title / company include & exclude, body exclusion phrases │
│ H-1B ─────── Company LCA data from MyVisaJobs (cached) │
│ Salary ───── Extracted from posting, H-1B data, description │
│ │
└────────────────────────────────────┬────────────────────────────────────────┘
│
▼
┌─────────────────────────────────────────────────────────────────────────────┐
│ JOB FEED │
│ │
│ Review ───── Dynamic filters, sorting, detail panel │
│ Decide ───── Save promising jobs, skip the rest, score with AI │
│ │
└────────────────────────────────────┬────────────────────────────────────────┘
│
▼
┌─────────────────────────────────────────────────────────────────────────────┐
│ AI RESUME SCORING │
│ │
│ Providers ── Claude API, Claude CLI, OpenAI, Ollama, OpenAI-compatible │
│ Depths ───── Light (scores only) or Full (report + keyword analysis) │
│ Multi ────── Score against multiple resumes, compare fit per role │
│ │
└────────────────────────────────────┬────────────────────────────────────────┘
│
▼
┌─────────────────────────────────────────────────────────────────────────────┐
│ RESUME + COVER LETTER │
│ │
│ Templates ── 8 resume + 8 cover-letter, auto-discovered (add your own) │
│ AI Tailor ── Rewrites resume bullets/keywords from the scoring report │
│ AI Letter ── Job-specific cover letters from resume + JD, voice presets │
│ Export ───── PDF via Playwright, page count indicator │
│ │
└────────────────────────────────────┬────────────────────────────────────────┘
│
▼
┌─────────────────────────────────────────────────────────────────────────────┐
│ TRACK │
│ │
│ Tracer ───── Unique links per resume/letter, tracks who opened them │
│ Gmail ────── Auto-detects responses, updates application status │
│ Telegram ─── Job alerts, daily digest, scrape health notifications │
│ │
└─────────────────────────────────────────────────────────────────────────────┘
| Feature | Description |
|---|---|
| Multi-Source Discovery | 6 scraping tiers: career pages (Playwright + 11 ATS), JobSpy (4 boards), LinkedIn Personal, Levels.fyi, Jobright.ai, Chrome Extension |
| AI Resume Scoring | Multi-provider (Claude, OpenAI, Ollama), light/full depth, per-resume comparison, keyword analysis, requirement mapping, ATS tips. Prompt caching on Anthropic cuts repeat-scoring cost ~50%. |
| Resume Builder | 8 templates (auto-discovered - add yours), AI tailoring per job, PDF export, tracer links to track opens (cover letters too) |
| Cover Letters | AI-generated per job, grounded in the paired resume + persona; editable voice presets, 8 templates, PDF export, prompt-cached generation |
| Smart Dedup | URL-hash dedup with configurable tracking-param stripping; content hash stored per job for cross-source matching |
| Job Feed | Filters, sorting, keyboard shortcuts (j/k/s/x/e), scoring reports, bulk operations, in-app job preview |
| Application Board | Kanban pipeline with drag-and-drop, status transition history |
| Chrome Extension | Passive LinkedIn capture + save any job from any page |
| Gmail Monitor | OAuth2 polling, auto-classifies responses, updates application status |
| Telegram Alerts | New job alerts, daily digest, scrape health, inline action buttons |
| H-1B Data | Company LCA lookups from MyVisaJobs, JD exclusion scanning |
| Scheduling | Cron-based: scraping, email checks, backups, cleanup, auto-reject |
| Dark Mode | Full Tailwind dark mode across all pages |
Note on the Job Feed preview pane: the detail panel renders the live job posting in an
iframe. Many career sites block being framed (viaX-Frame-Options/ CSPframe-ancestors, or cross-origin scripts that fail when embedded), so the in-app preview works best with a browser extension that strips frame-blocking headers (e.g. an "ignore X-Frame-Options" extension). Without one, some postings show blank — use the "Open" button to view them in a new tab. Applied jobs fall back to a cached snapshot that always renders.
git clone https://github.com/vesaias/JobNavigator.git
cd JobNavigator
cp .env.example .env
# Edit .env if needed (optional — API keys can be set from dashboard)
docker compose up --build -dOpen http://localhost. On first run, click "Sign In" with a blank API key to proceed. Set a real key from Settings > General once you've accessed the dashboard.
First steps:
- Settings > AI tab — configure your LLM provider and API key
- Companies — activate a few seed companies or add your own
- Searches — configure a keyword search or activate LinkedIn Personal
- Resumes — create your resume (or import an existing PDF); it powers AI scoring
chrome://extensions/→ Developer mode → Load unpacked → selectextension/- Toggle LinkedIn capture on, browse job collections
- Click "Send to JobNavigator" to import
Telegram — Create bot via @BotFather, set token in .env, enter chat ID in Settings.
Gmail — Run python backend/gmail_oauth_setup.py, set OAuth credentials in .env.
| Layer | Technology |
|---|---|
| Backend | Python 3.12, FastAPI, SQLAlchemy, APScheduler, Playwright |
| Frontend | React 18, Tailwind CSS, Vite, Recharts |
| Database | PostgreSQL 16 |
| Infrastructure | Docker Compose, Caddy, nginx |
| AI | Anthropic SDK, OpenAI SDK, Ollama, Claude Code CLI |
| Extension | Chrome Manifest V3 |
PRs welcome — see CONTRIBUTING.md. Easy wins: new ATS scrapers, new resume templates, UI polish.
Found a vulnerability? See SECURITY.md. Please don't open public issues for security bugs.
JobNavigator is self-hosted — NOT a hosted service. Your resume, job data, and credentials stay on your machine. Data is sent only to the AI provider you configure. We do not collect, store, or have access to any of your data.
Personal and educational use only. Not affiliated with LinkedIn, Indeed, Jobright.ai, or any job platform. Some scraping features are disabled by default and require explicit opt-in. You are responsible for complying with the Terms of Service of any platform you interact with. See LEGAL_DISCLAIMER.md.