ConvertIn

LinkedIn Profile → Polished One-Page CV in 60 Seconds

---

Overview

ConvertIn is a self-hosted web app that turns a LinkedIn PDF export into a tailored, ATS-friendly one-page CV. Upload your profile, edit the extracted data, pick a template and accent color, and download a professional PDF — no sign-up, no login, no cloud dependencies.

An LLM (OpenAI, DeepSeek, or any OpenAI-compatible API) reads your LinkedIn PDF, extracts your experience, education, and skills, then generates a CV optimized for your target role and language. Four built-in templates auto-scale fonts to always fit one page.

No accounts. Profiles are accessed by UUID token in the URL and expire after 24 hours.

---

Features

• LinkedIn PDF Parsing — Upload your exported LinkedIn PDF; an LLM extracts name, headline, experience, education, skills, certifications, and languages.
• Profile Editor — Fine-tune every field before generating your CV. Update email, phone, location, and LinkedIn URL separately from the main profile.
• AI CV Generation — Specify a target role and language; the LLM curates the most relevant experience and writes tailored bullet points.
• 4 CV Templates — Modern (two-column with accent color), Professional (single-column), Harvard (classic format), and Simple (minimal text-focused).
• Smart Font Scaling — All templates auto-scale text (0.82×–1.55×) based on content density to guarantee a one-page fit.
• Photo Upload — Add a headshot (PNG/JPEG/WebP, up to 5 MB) stored locally and embedded in the PDF.
• Configurable LLM — Works with OpenAI, DeepSeek, or any OpenAI-compatible endpoint via LLM_BASE_URL / LLM_API_KEY / LLM_MODEL.
• Zero Cloud Dependencies — SQLite database, local file storage, no external services required beyond the LLM API.
• Privacy-First — No auth, no accounts. Profiles expire after 24 hours and are cleaned up automatically.

---

Tech Stack

Layer	Technology
Framework	Next.js 16 (App Router, Server Components, API Routes)
Language	TypeScript 5
Database	SQLite via better-sqlite3 (WAL mode, auto-migration)
LLM	OpenAI SDK (configurable provider)
PDF	@react-pdf/renderer (server-side, 4 templates)
Styling	Tailwind CSS v4 + shadcn/ui patterns
Validation	Zod
Testing	Vitest + jsdom

---

Getting Started

Prerequisites

• Node.js 20+ (tested on 26)
• Build tools for native modules — Xcode Command Line Tools on macOS, build-essential on Linux
• An LLM API key — OpenAI, DeepSeek, or any OpenAI-compatible provider

Installation

1. Clone the repository

   git clone https://github.com/AzuriteAzul/convertin.git
   cd convertin

2. Install dependencies

   npm install

3. Configure environment variables

   cp .env.example .env.local

   Edit .env.local with your LLM credentials:

   LLM_API_KEY=sk-your-api-key
   LLM_BASE_URL=https://api.openai.com/v1
   LLM_MODEL=gpt-4o

   For DeepSeek:

   LLM_API_KEY=your-deepseek-key
   LLM_BASE_URL=https://api.deepseek.com/v1
   LLM_MODEL=deepseek-v4-pro

4. Start the development server

   npm run dev

   The app will be available at http://localhost:3000.

---

Available Scripts

Script	Description
npm run dev	Start the Next.js development server
npm run build	Production build (always verify with this)
npm run start	Start the production server
npm run lint	ESLint with next/core-web-vitals + next/typescript
npm run test	Vitest single run
npm run test:watch	Vitest watch mode
npm run test:coverage	Vitest with coverage

---

Project Structure

convertin/
├── public/fonts/                # Inter font files for PDF templates
├── src/
│   ├── app/
│   │   ├── api/
│   │   │   ├── profiles/        # Profile CRUD + PDF upload
│   │   │   ├── cv/              # CV generation + PDF download
│   │   │   └── photos/          # Photo serving from local disk
│   │   ├── generate/            # Upload & edit page
│   │   ├── preview/             # CV preview page
│   │   ├── install/             # Install prompt page
│   │   ├── layout.tsx           # Root layout (Geist font, metadata)
│   │   └── page.tsx             # Landing page
│   ├── components/              # React components
│   │   ├── ProfileEditor.tsx    # Edit extracted profile data
│   │   ├── TemplateSelector.tsx # Choose CV template
│   │   ├── ColorPicker.tsx      # Accent color picker
│   │   ├── LanguageSelector.tsx # CV output language
│   │   ├── RoleSelector.tsx     # Target role input
│   │   ├── PhotoUploader.tsx    # Headshot upload
│   │   ├── CvTemplateModern.tsx
│   │   ├── CvTemplateProfessional.tsx
│   │   ├── CvTemplateHarvard.tsx
│   │   └── CvTemplateSimple.tsx
│   ├── lib/
│   │   ├── db.ts               # SQLite database (better-sqlite3, WAL)
│   │   ├── llm.ts               # LLM client (configurable provider)
│   │   ├── pdf.ts               # PDF rendering + font scaling
│   │   ├── storage.ts           # Local filesystem photo storage
│   │   ├── cv-helpers.ts        # Shared CV template utilities
│   │   └── utils.ts             # dbToProfile, cn(), generateToken
│   └── types/                   # Zod schemas & TypeScript types
├── test/                        # Vitest tests
├── data/                         # SQLite DB + photo uploads (gitignored)
├── .env.example                 # Environment variable template
├── next.config.ts               # Next.js config (serverExternalPackages)
├── vitest.config.ts             # Vitest configuration
└── AGENTS.md                    # AI agent instructions

---

CV Templates

Template	Layout	Description
Modern	Two-column	Clean layout with configurable accent color sidebar
Professional	Single-column	Traditional format, emphasis on content hierarchy
Harvard	Single-column	Classic Harvard Business School résumé format
Simple	Single-column	Minimal, text-focused layout

All templates use smart font scaling — calculateFontScale() in src/lib/pdf.ts adjusts text size from 0.82× to 1.55× based on content density, ensuring every CV fits on exactly one page.

---

Key Conventions

• No authentication. Profiles are accessed by UUID session token in query params. Share the URL, share access.
• 24-hour expiry. Profiles and their data are automatically deleted after 24 hours. Cleanup runs on DB init.
• Snake_case in DB, camelCase in TypeScript. Always use dbToProfile() from src/lib/utils.ts at the database boundary.
• raw_data JSON column. Email, phone, and LinkedIn URL are stored in a JSON column on profiles, not as separate columns. Always extract via dbToProfile().
• Contact override. POST /api/cv/generate overwrites the LLM's contact section with the user's actual email/phone/location/linkedinUrl to prevent stale data.
• PDF uses React.createElement. JSX doesn't work server-side with @react-pdf/renderer. Templates use React.createElement directly.
• serverExternalPackages. Both pdfjs-dist and better-sqlite3 must stay in next.config.ts's serverExternalPackages — removing either breaks the build.
• Singleton clients. LLM and DB use module-level singletons. resetLlmClient() and resetDb() exports exist for tests.
• Photo URLs are stored as /api/photos/<token> (no file extension). The photo route finds files by token prefix.
• Headline trimming. The PDF route strips text after @, |, -, —, · and locale-specific employer suffixes to extract just the job title.

---

Deployment

ConvertIn is designed for self-hosting. Since it uses SQLite and local file storage, it runs on any Node.js server:

1. Push your code to GitHub.
2. Clone on your server and install dependencies.
3. Configure .env.local with your LLM credentials.
4. Run npm run build && npm run start.

Note: The data/ directory (SQLite database + uploaded photos) must persist between deployments. Do not mount it as ephemeral storage.

For Docker or reverse-proxy setups, ensure the data/uploads/photos/ directory exists and is writable by the Node process.

---

Contributing

Contributions are welcome! Please follow these steps:

1. Fork the repository.
2. Create a feature branch (git checkout -b feature/your-feature).
3. Make your changes and ensure npm run lint and npm run test pass.
4. Open a pull request with a clear description of what changed and why.

If you are adding new CV templates, remember to:

• Add the template component in src/components/.
• Register it in src/lib/pdf.ts's template map.
• Add shared utilities to src/lib/cv-helpers.ts (don't duplicate parseBullets or hasItems locally).
• Update the CvJsonSchema metadata type in src/types/index.ts if the template introduces new options.

---

License

This project is licensed under the MIT License. See LICENSE for details.

---

Built for job seekers who deserve better than LinkedIn's default PDF.