Smart Job Seeker
An intelligent job-search and application-management platform: create a detailed
profile once, automatically match it against job listings, and manage your entire
application process with AI assistance.
This repository is the open-source web application — the SvelteKit app,
database, matching engine, and AI features. It is one part of a larger product;
the actual job-scraping engine is a separate, closed-source component (see
Project components below).
Project components
Smart Job Seeker is built from several components. This repo is the open one:
| Component | Open? | What it does |
|---|
| Web app (this repository) | ✅ open source (GPLv3) | SvelteKit app + PostgreSQL: profiles & portfolio, job and application management, AI writing features, profile↔job matching, and the dashboard that schedules and queues scraping runs. |
| Device client | ✅ open source — sjs-browser, sjs-desktop | Runs a real browser on your own device and connects to the SJS servers to receive navigation instructions, which it executes locally against the job sites (CDP bridge, stealth, live view). sjs-browser is a headless/Docker build; sjs-desktop is a Tauri desktop app. Both are separate public repos. |
| Scraping / extraction worker | 🔒 closed source | Runs on the SJS servers. Decides how to navigate each job site, streams those instructions to your device client, and performs the LLM-based job extraction on the pages it returns. |
In other words: this app orchestrates scraping — it manages search tasks,
enqueues runs, ingests results, and matches them against profiles. The browser
that actually visits job sites runs on your own device via the open-source
sjs-browser /
sjs-desktop clients; those clients
connect to the SJS servers, where the closed-source worker decides how to
navigate each site and extracts the job data with an LLM. That closed worker is
not included here.
What that means if you run this repo on its own:
- ✅ Profiles & portfolio, application tracking, AI writing, JSON-Resume/PDF
export, and profile↔job matching all work against data already in the DB.
- ⚠️ Live job scraping is not available without the closed-source worker —
scraping runs you enqueue will sit in the queue with nothing to process them.
Features
Profile & portfolio
- Comprehensive profiles — work experience, education, skills, projects
- Personal portfolio — responsive showcase of your professional background
- Data export — multiple formats, including JSON Resume
with PDF output
Matching & applications
- Profile↔job matching — scores jobs against your profile, with an admin
dashboard and an optional semantic skill-matching layer
- Application tracking — organize jobs, interviews, and follow-ups
- AI-powered writing — generate cover letters, follow-ups, and thank-you notes
- Interview prep — AI-generated answers to application questions
- Activity logging and file management
Technical highlights
- Multi-provider LLM integration (e.g. Groq, Anthropic, OpenAI, Gemini, DeepSeek)
- Background job queues (scrape / match / re-scrape) backed by Redis
- Type-safe data layer with Drizzle ORM
Quick start
Prerequisites
- Docker & Docker Compose (the dev environment runs entirely in containers)
- Node.js — only needed for host-side tooling; the version is pinned in
.nvmrc
Run
git clone <repository-url>
cd smart-job-seeker
# Configure environment
cp .env.example .env
# Edit .env with your configuration (DB, JWT secret, LLM API keys, …)
# Start the dev environment (app + database + adminer)
npm run dev
This starts:
On first run with an empty database, Drizzle migrations are applied
automatically. No sample dataset is bundled with this repository.
Architecture
Tech stack
- Frontend: SvelteKit 5, Svelte 5, TypeScript, Tailwind CSS
- Backend: Node.js, SvelteKit server routes, Drizzle ORM
- Database: PostgreSQL (+ Adminer for inspection)
- Queues: Redis-backed background jobs
- AI: multi-provider LLM abstraction
- Dev/Ops: Docker Compose, Vitest
Key modules (in this repo)
src/lib/server/db/ — Drizzle schema and data access (schema.ts)src/lib/server/job/ — matching logic (match-utils.ts, matcher-state.ts,
match-trigger.ts)src/lib/server/queue/ — background job queues that orchestrate scraping and
matching (scraper-queue.ts, match-queue.ts, rescrape-queue.ts)src/lib/server/ai-chat/ — AI prompt templates and writing featuressrc/lib/server/profile/ — profile export / collected_data snapshottingsrc/routes/ — SvelteKit app and API routes
The browser that visits job sites runs on your own device via the open-source
sjs-browser /
sjs-desktop clients; the
closed-source worker on the SJS servers tells it how to navigate and extracts
the job data with an LLM. Neither is in this repository.
Development
# Development
npm run dev # start app + database + adminer
npm run dev:reset # reset the database
npm run docker:cli # shell into the app container
# Database (Drizzle)
npx drizzle-kit push # apply schema changes in dev
npm run docker:db:backup # back up the local database
# Quality
npm run check # type checking
npm run test # unit tests (Vitest)
npx deno fmt # format
See the docs directory for more:
License
GNU General Public License v3.0 — see LICENSE.
About
Built by Rik Wanders, Senior Full Stack Developer.
