A job search dashboard for engineers who'd rather edit TypeScript than drag cards around in Trello.
No magic. No "AI finds your dream job" nonsense. You do the searching, you make the decisions, this gives you a clean interface to track it all and stop losing leads in browser tabs and half-finished spreadsheets.
You're a developer looking for work. You know what you're worth, you know what stack you want, and you don't need a chatbot to tell you what jobs to apply for. What you need is:
- A place to dump search results without them disappearing into the void
- A way to rank jobs by how well they actually fit your profile — not by Indeed's algorithm
- A system to track which search terms work and which are a waste of time
- A dashboard that loads instantly and doesn't ask you to sign up for anything
If you want an app that auto-applies to 500 jobs and writes your cover letter, this isn't it. If you want a sharp tool for managing a focused, intentional job search — keep reading.
- Priority Tiers (P1/P2/P3) — Every job gets a tier. P1 = strong fit, apply now. P2 = worth a look. P3 = you could do it. Skip = don't bother. Focus your energy where it matters.
- Company Targets — Pick a company, see all their open roles scored by confidence %, sorted by fit. Includes company intel (salary averages, CEO approval, interview difficulty) and your strategic angle for each target.
- Confidence Scoring — Auto-calculated 0-100% match based on title keywords, seniority, role type, salary range, and remote availability. Customize the weights in
targets.tsto match your own profile. - Title Analytics — Which job title searches actually produce relevant results? Track verdicts (Works / Sometimes / Never) so you stop repeating dead-end queries.
- Inline JDs — Job descriptions render inside the app when available. No tab-switching to Indeed.
- Application Pipeline — Kanban board and table view: Saved → Applied → Screening → Interview → Offer → Accepted / Rejected.
- Deploy-Ready Tracking — See which applications have cover letter + portfolio packages built and ready to submit.
- Demo/Live Toggle — Switch between demo data (for screenshots) and your real data with one click. Persists across reloads via localStorage.
- Indeed Saved Jobs Import — If you can grab the JSON from Indeed's saved jobs API, the data maps directly into the search results format with metadata like
normalizedJobTitle,easyApply, andexpiredstatus.
Company Targets — confidence scoring + company intel
Title Analytics — track which searches work vs waste time
Profile — strengths, targets, talking points
| Tier | Meaning | Color |
|---|---|---|
| P1 | Strong fit — apply now | Green |
| P2 | Almost — worth a look | Amber |
| P3 | You can do this | Blue |
| Skip | Not a fit | Gray |
| Layer | Technology |
|---|---|
| Framework | SvelteKit with Svelte 5 (runes: $state, $derived, $props) |
| Styling | Tailwind CSS v4 via @tailwindcss/vite |
| Language | TypeScript — strict, fully typed data layer |
| Runtime | Bun — fast installs, fast dev server |
| Data | Plain TypeScript files — no database, no auth, no backend |
| Theme | Dark mode, #39FF14 neon green accent |
git clone https://github.com/NooRotic/JobTrackr.git
cd JobTrackr
bun install
bun run devOpen http://localhost:5173. The app ships with example data so you can see how everything works.
All data lives in src/lib/data/. Plain TypeScript files with typed interfaces. No database, no migration, no setup wizard. Open the file, edit the data, save, and the dev server hot-reloads.
| File | What it does |
|---|---|
searches.ts |
Your saved job searches with results and priority tiers |
targets.ts |
Company targets, job title categories, confidence scoring algorithm |
titleAnalytics.ts |
Job title effectiveness tracking |
applications.ts |
Application pipeline entries |
companies.ts |
Company research cards |
activity.ts |
Activity feed |
profile.ts |
Career profile data |
types.ts |
All TypeScript interfaces — extend as needed |
Replace the example data with your own. That's it.
Your real job search data should not live in a public repo. The .gitignore already excludes src/lib/data/personal/. Keep your real data files there — the app loads personal data by default and falls back to demo data. Toggle between them with the sidebar button or ?demo=true URL param. Your preference persists in localStorage across reloads.
JobTrackr was built alongside Claude Code with the Indeed MCP plugin. This is entirely optional — the dashboard works fine with manually entered data — but if you have this setup, the workflow gets significantly faster:
What Claude Code + Indeed MCP enables:
search_jobs— Run Indeed searches by keyword, location, and job type. Results map directly intosearches.tsformat.get_job_details— Pull full job descriptions by job ID. Populates thejobDescriptionfield for inline JD viewing.get_company_data— Pull company ratings, salary averages, CEO approval, interview difficulty. Populates company target cards.get_resume— Verify how Indeed parsed your uploaded resume.
The workflow:
- Tell Claude: "Search YouTube for Senior Software Engineer roles"
- Claude runs the searches, scores results, assigns priority tiers
- Claude writes the results directly into your data files
- Dashboard hot-reloads with new results
What it does NOT do:
- Apply for jobs on your behalf
- Write cover letters or resumes
- Make decisions about which jobs to pursue
- Anything automatically — you ask, it executes, you decide
This is a power tool, not an autopilot.
src/
lib/
components/ # StatusBadge, FitBadge, PriorityBadge, DeployBadge, StatCard
data/ # All typed data files (example data ships with repo)
personal/ # Your real data (gitignored)
routes/
+layout.svelte # Root layout with sidebar nav + demo/live toggle
+page.svelte # Dashboard home (stats, leads, pipeline, activity)
searches/ # Priority-tiered job board (P1/P2/P3)
targets/ # Company-targeted search with confidence scoring
applications/ # Kanban board + table view with add/edit
companies/ # Company research card grid
analytics/ # Job title effectiveness tracking
profile/ # Career profile with talking points
bun run build
bun run previewDeploy to Vercel, Cloudflare Pages, GitHub Pages, or any static host with the appropriate SvelteKit adapter.
PRs welcome. This is intentionally minimal — no database, no auth, no server-side rendering of dynamic data. The goal is a fast, portable dashboard that any developer can fork and run locally in under a minute.
If you're adding features, keep them opt-in and data-file driven. The core value is simplicity.
MIT — see LICENSE.