ApplyKit

Self-hosted, local-first CV and cover letter generator powered by AI. Your data stays on your machine — no cloud, no account, no subscription.

⚠️ Self-hosted only. ApplyKit is designed to run locally or on your own server. It has no authentication — do not expose it to the public internet without adding your own auth layer.

Features

• Multi-profile support — create separate profiles for different roles (e.g. "Software Engineer", "Product Manager"), each with its own history and color
• AI CV generation — rewrites bullet points and summary to be ATS-optimized for a specific job description
• AI cover letter generation — writes a tailored cover letter from your profile + job description in seconds
• CV import — paste or upload an existing PDF/DOCX and AI extracts your profile automatically
• Job URL scraper — paste a job posting URL and AI extracts the description
• Fit analysis — see how well your profile matches a job description with match score, strengths, gaps, and red flags
• Smart Apply — paste a URL, auto-parse job details, and generate a tailored CV + cover letter in one flow
• Job application tracker — Kanban board for tracking applications through Applied → Interviewing → Offer → Rejected
• History — every generated CV and cover letter is saved, browsable, and filterable by profile
• PDF export — download as PDF or use browser print (A4 format)
• LLM-agnostic — works with Gemini, OpenAI, Anthropic, or any local model via Ollama
• Works without an API key — CV generation falls back to your raw profile data if no LLM is configured
• Dark mode — full dark theme support

How It Works

Browser (localhost:5173)
       │
       ▼
  SvelteKit Frontend
  - Profile editor (multi-profile)
  - CV preview & export
  - Cover letter editor
  - Fit analysis
  - Job tracker (Kanban)
  - History browser
       │  HTTP (REST API)
       ▼
  FastAPI Backend (localhost:8000)
  - Profile CRUD (SQLite, multi-profile)
  - CV import (PDF/DOCX/text → LLM extraction)
  - ATS CV enhancement (LLM)
  - Cover letter generation (LLM)
  - Fit score analysis (LLM)
  - Job URL scraping (Jina + Crawl4AI)
  - PDF export (WeasyPrint)
  - Generation history
       │
       ▼
  LiteLLM → any LLM provider (Gemini, OpenAI, Anthropic, Ollama...)

Stack

Prerequisites

• uv — Python package manager
• Bun — JavaScript runtime
• WeasyPrint system dependencies (for PDF generation):
  • Ubuntu/Debian: apt-get install libcairo2 libpango-1.0-0 libgdk-pixbuf2.0-0 libffi7 shared-mime-info
  • macOS: brew install cairo pango gdk-pixbuf
  • Windows: Included in the WeasyPrint pip package
• An LLM API key (optional — required for AI features):
  • Google AI Studio for Gemini (recommended, has a generous free tier)
  • OpenAI, Anthropic, or any LiteLLM-supported provider
  • Or Ollama for fully local, offline usage

Quick Start

Option A — Docker (recommended)

git clone https://github.com/wihlarkop/applykit.git
cd applykit
docker compose up --build

Open http://localhost:3000 — your data is stored in a Docker volume and persists across restarts.

Option B — Manual

# 1. Clone
git clone https://github.com/wihlarkop/applykit.git
cd applykit

# 2. Configure
cp backend/.env.example backend/.env

# 3. Install dependencies
make install

# 4. Run database migrations
make migrate

# 5. Start (two separate terminals)
make backend    # http://localhost:8000
make frontend   # http://localhost:5173

Open http://localhost:5173 — you'll be guided through setup on first launch.

Docker

Default setup

docker compose up --build

• Frontend: http://localhost:3000
• Backend API: http://localhost:8000
• SQLite data: persisted in a named Docker volume (applykit_applykit-data)

Deploying to a remote server

If you're running on a server instead of localhost, set VITE_API_BASE_URL to your domain before building:

VITE_API_BASE_URL=https://api.yourdomain.com/api docker compose up --build

Or edit the args block in docker-compose.yml:

args:
  VITE_API_BASE_URL: https://api.yourdomain.com/api

VITE_API_BASE_URL is baked into the frontend at build time — the browser uses it to reach the backend. It must be publicly reachable from the user's machine.

Backing up your data

SQLite is stored in a Docker volume. To export it:

docker run --rm \
  -v applykit_applykit-data:/data \
  -v $(pwd):/backup \
  alpine tar czf /backup/applykit-backup.tar.gz /data

Useful Docker commands

docker compose up --build        # Build and start
docker compose up -d             # Start in background
docker compose down              # Stop
docker compose logs -f backend   # Follow backend logs
docker compose exec backend uv run alembic upgrade head  # Run migrations manually

Setup (Manual)

1. Clone the repository

git clone https://github.com/your-username/applykit.git
cd applykit

2. Configure the backend

cd backend
cp .env.example .env

3. Install backend dependencies

uv sync

4. Run database migrations

uv run alembic upgrade head

5. Start the backend

uv run main.py
# API: http://localhost:8000
# Swagger UI: http://localhost:8000/docs

6. Install and start the frontend

cd ../frontend
bun install
bun run dev
# Frontend: http://localhost:5173

Makefile Commands

make install        # Install all dependencies (backend + frontend)
make migrate        # Run database migrations
make backend        # Start backend server (http://localhost:8000)
make frontend       # Start frontend dev server (http://localhost:5173)
make lint           # Lint frontend TypeScript/Svelte
make migrate-new MSG="description"  # Create a new migration
make migrate-down                   # Roll back the last migration
make help           # Show all commands

Configuration

Database

Edit backend/.env to change the database path:

DATABASE_URL=sqlite:///./applykit.db

SQLite is the default and requires no additional setup. PostgreSQL is on the roadmap.

LLM Settings

LLM configuration (provider, API key, model) is managed via the Settings page in the UI — no need to edit .env manually.

Click the Settings icon (gear) in the top navigation to connect a provider. You can connect multiple providers and switch between them at any time.

No API key? The app still works. CV generation falls back to your raw profile data without AI enhancement. Import, cover letter generation, and fit analysis require an LLM to be configured.

Usage

1. Create a profile

On first launch you'll be guided through setup. Fill in:

• Personal info (name, email, location, LinkedIn, GitHub)
• Work experience with bullet points
• Education, skills, projects, certifications

Or use AI Sync (sparkle button on the profile page) to upload an existing CV and auto-fill everything instantly.

2. Generate a CV

1. Go to Generate CV
2. Optionally paste a job description — the AI will tailor bullet points to match its keywords
3. Click Generate ATS CV
4. Preview, then Download PDF or Print

3. Write a cover letter

1. Go to Cover Letter
2. Fill in company name (optional), job description, and any emphasis notes
3. Click Write Cover Letter
4. Copy, download as PDF, or print

4. Analyze job fit

Paste a job description in the Cover Letter page and click Analyze Fit to see:

• Match score (0–100%)
• Strengths and gaps
• Missing keywords
• Red flags
• Suggested emphasis for your cover letter

5. Smart Apply

1. Go to Smart Apply
2. Paste a job posting URL — ApplyKit scrapes it automatically
3. Review the extracted job details
4. Generate a tailored CV + cover letter in one click

6. Track applications

Go to Tracker to add jobs, drag cards between stages, and link generated CVs and cover letters to each application.

7. Browse history

Go to History to see every generated CV and cover letter. Filter by profile, search, sort by match score, and preview or re-download any entry.

Smart Apply — Supported Job Boards

ATS Platforms (Direct API — Fastest)

Generic Websites

For boards without direct API support, Smart Apply uses Jina to scrape the page and an LLM to extract structured fields. This works on most job sites.

Security

ApplyKit has no built-in authentication. It is designed to run:

• On localhost for personal use (default)
• On a private server with network-level access control

Do not expose ApplyKit to the public internet without putting an auth proxy (e.g. Authelia, Nginx basic auth, Cloudflare Access) in front of it.

All LLM API keys are stored in your local SQLite database and never leave your machine.

Roadmap

Items marked ✅ are shipped. Items marked 📋 are planned.

Core App

AI Providers

UX & Polish

Future Ideas

Contributing

Contributions are welcome! Here's how to get started:

1. Fork the repository and create a branch: git checkout -b feat/your-feature
2. Set up locally following the Quick Start guide above
3. Install pre-commit hooks (one-time setup):

   bun install
   bun x lefthook install
   

   This wires up automatic linting and formatting on every git commit. Requires uv to be installed.
4. Make your changes — keep PRs focused on a single feature or fix
5. Test your changes manually (run both backend and frontend)
6. Submit a PR with a clear description of what changed and why

Good first issues

• Adding support for a new ATS platform in Smart Apply
• Improving PDF export styling
• Adding keyboard shortcuts
• Translating the UI

Guidelines

• Keep the self-hosted, local-first philosophy — no mandatory cloud services
• Don't add authentication logic to the core app (out of scope)
• Follow existing code style: Svelte 5 runes, Tailwind CSS, FastAPI patterns
• For large features, open an issue first to discuss the approach

License

MIT — see LICENSE