Cover Letter Generator

Project Overview

This project is a full-stack web application that generates personalized, job-specific cover letters by combining a candidate’s resume with a job description using an AI language model. It automates the repetitive and time-consuming process of writing tailored cover letters, enabling users to generate, edit, and export high-quality documents in minutes.

Problem Addressed

Writing customized cover letters for multiple job applications is repetitive and inefficient.
Most applicants reuse generic templates, which reduces quality and relevance.

This application automates the personalization process by dynamically aligning resume content with job descriptions while still allowing users to review and edit the final output.

Generation Flow

1. User uploads a resume file (PDF / DOC / DOCX)
2. User provides company name, position title, and job description
3. Optional profile metadata is included for personalization
4. Form data is persisted in sessionStorage to prevent data loss
5. A serverless API route:
   • Extracts resume text
   • Constructs a structured AI prompt
   • Calls the OpenAI API
6. Generated cover letter is returned to the client
7. User can edit, copy, export as PDF, or save locally

AI & Prompt Engineering

• Uses GPT-4o-mini for cost-efficient, high-quality text generation
• Prompt structure consists of:
  • System prompt defining role, tone, formatting, and constraints
  • User prompt combining resume content, job description, profile data, and current date
• Explicit instructions enforce:
  • No placeholders
  • Consistent formatting
  • Target length (300–400 words)
• Prompt templates are:
  • Editable by the user
  • Persisted in localStorage
  • Resettable to defaults

Architecture & State Management

• Stateless serverless backend with a single REST-style API endpoint
• Client-side state layered across:
  • React component state (UI & loading states)
  • sessionStorage for form persistence across reloads
  • localStorage for user profile, prompt templates, and generation history
• Resume files are:
  • Converted to Base64 for session persistence
  • Reconstructed as File objects when restoring state

PDF & Output Handling

• Custom PDF generation using jsPDF
• Implements:
  • Paragraph-aware text splitting
  • Pagination logic for multi-page documents
  • Dynamic file naming using candidate and company metadata
• Generated content remains editable before export, enabling manual refinement

Error Handling & UX

• Server-side validation for required fields
• Graceful handling of:
  • OpenAI API failures
  • File processing errors
• Clear user feedback via:
  • Disabled submit states
  • Inline error messages
  • Toast notifications
• No page reloads during generation or editing flow

Security & Cost Considerations

• OpenAI API key stored exclusively in environment variables
• No sensitive data persisted on the server
• All user data remains client-side
• Cost control via:
  • Token limits
  • Model selection optimized for price/performance

Architectural Summary

• Full-stack Next.js application using App Router and TypeScript
• Serverless API route handling file uploads and AI requests
• Context-aware prompt construction from multiple user inputs
• Client-side persistence strategy combining localStorage and sessionStorage
• Custom document generation pipeline with export and history support
• Designed for practical UX, low operational overhead, and easy deployment