Skip to content

sweady-labs/JobMatchAI

BranchesTags

Repository files navigation

JobMatch AI - Application Letter Agent

An AI-powered job application automation system that generates tailored, professional cover letters by combining your competence profile with job postings.

Application workflow overview

Table of Contents

Quick Start (15 minutes)

# 1. Clone the repository
git clone https://github.com/sweady1989/JobMatchAI
cd JobMatchAI

# 2. Copy example files to create your configuration
cp 3_latex/user_info.example.yml 3_latex/user_info.yml

# 3. Edit user_info.yml with your personal details
# (Use your favorite text editor)

# 4. Start n8n (requires Docker)
cd 4_n8n
docker compose up -d

# 5. Open http://localhost:5678 in your browser
# 6. Import the workflow: 4_n8n/application-letter-agent.json
# 7. Add your OpenAI API credentials in n8n
# 8. Add a job posting to candidates/example_user/0_inbox-jobs/
# 9. Run the workflow!

Project Structure

JobMatchAI/
├── 0_queries&prompts/           # GPT prompts and LinkedIn search queries
│   ├── prompts/                 # AI prompts for letter generation
│   │   └── 3_last_update_ATS_optimized_prompt.md  # Main prompt
│   └── queries_linkedin/        # Job search query templates
│       └── search_queries_template.md
│
├── 3_latex/                     # LaTeX document system
│   ├── src/
│   │   ├── templates/           # Cover letter and CV templates
│   │   ├── content/             # Generated .md files go here for processing
│   │   └── applications/        # Converted .tex files
│   ├── figures/                 # Images (profile pics, signatures)
│   ├── output/                  # Final PDFs appear here
│   ├── build/                   # Build artifacts and logs
│   ├── parse_md_to_tex.py       # Markdown → LaTeX converter
│   ├── user_info.example.yml    # TEMPLATE: Copy to user_info.yml
│   └── Makefile                 # Build commands
│
├── 4_n8n/                       # Workflow automation
│   ├── docker-compose.yml       # Docker configuration
│   ├── application-letter-agent.json  # Main workflow (import this!)
│   ├── IMPROVEMENT_GUIDE.md     # Technical documentation
│   └── application_letter_cost_analysis.md  # Cost tracking guide
│
├── candidates/                  # Your application data
│   └── example_user/            # Example candidate folder (copy for yourself)
│       ├── 0_inbox-jobs/        # Job postings go here
│       │   ├── 0_template_jobs.md           # Template for new jobs
│       │   └── 2025-01-15/                  # Date-organized folders
│       │       └── 2025-01-15_company_role.md
│       ├── 1_profile/           # Your competence profile
│       │   └── competence-profile.md        # EDIT THIS with your info
│       ├── 2_applications-mds/  # Generated letters appear here
│       │   └── 2025-01-15/
│       │       └── 2025-01-15_company_role_ENG.md
│       ├── 3_job_analysis/      # Job match analysis (optional)
│       └── .latest_job.json     # Points to current job folder
│
├── 98_testfiles/                # Test and experiment files
├── Makefile                     # Top-level build commands
├── normalize_mds.py             # Filename cleanup utility
└── README.md                    # This file

Detailed Setup Guide

Prerequisites

Install these tools before starting:

Tool macOS Windows Purpose
Git brew install git Download Version control
Docker Desktop brew install --cask docker Download Runs n8n workflow engine
Python 3 brew install python Download Runs conversion scripts
LaTeX (XeLaTeX) brew install --cask mactex MiKTeX Compiles PDF documents

After installing, restart your terminal.

Clone and Configure

# Clone the repository
git clone https://github.com/sweady1989/JobMatchAI
cd JobMatchAI

# Create your personal configuration (this file is gitignored)
cp 3_latex/user_info.example.yml 3_latex/user_info.yml

Now edit 3_latex/user_info.yml with your real information:

- id: yourname                    # Used to select your profile
  full_name: Your Full Name
  email: [email protected]
  phone: "+49 123 456 7890"
  job_title: Your Target Role
  linkedin_url: https://www.linkedin.com/in/yourprofile/
  location: Your City, Country
  # ... fill in all fields

Set Up Your Profile

Your competence profile is the key to great application letters. The AI uses this to match your skills to job requirements.

  1. Navigate to candidates/example_user/1_profile/competence-profile.md
  2. Study the example structure - it shows you exactly what information to include
  3. Replace with your own information:
    • Professional summary - 2-3 sentences about who you are
    • Work experience - Include achievements with metrics ("Reduced costs by 30%")
    • Education - Degrees, certifications, relevant courses
    • Skills - Technical skills, tools, methodologies
    • Notable projects - Highlight your best work

Tips for a strong profile:

  • Be specific: "Reduced deployment time by 70%" beats "Improved efficiency"
  • Use keywords from your target industry
  • Include both technical skills and soft skills
  • Update regularly as you gain new experience

Configure n8n Workflow

n8n is the workflow automation engine that orchestrates the AI letter generation.

# Start n8n
cd 4_n8n
docker compose up -d
  1. Open http://localhost:5678 in your browser
  2. Log in with credentials from docker-compose.yml (change the default password!)
  3. Click Import workflow → Select application-letter-agent.json
  4. Add your OpenAI API key:
    • Click on the "create prompt-> send to model" node
    • Click on "Credentials" → "Create new credential"
    • Select "OpenAI API"
    • Enter your API key from https://platform.openai.com/api-keys
  5. Update file paths in the workflow if you created a custom candidate folder:
    • path competence profile node: Update the path to your profile
    • reading json from "create-daily-folders"-workflow node: Update the path to your .latest_job.json

Add a Job Posting

  1. Create a date folder for today:

    mkdir -p candidates/example_user/0_inbox-jobs/$(date +%Y-%m-%d)

  2. Copy the template:

    cp candidates/example_user/0_inbox-jobs/0_template_jobs.md \
   "candidates/example_user/0_inbox-jobs/$(date +%Y-%m-%d)/$(date +%Y-%m-%d)_company_jobtitle.md"

  3. Open the file and paste the job description

  4. Fill in the metadata (company, location, salary if known)

  5. Save the file

Generate Your Letter

  1. Update .latest_job.json with the path to your job folder:

    {
  "jobs": [
    { "jobFilePath": "/workspace/jobs/candidates/example_user/0_inbox-jobs/2025-01-15" }
  ]
}

  2. In n8n, click Execute Workflow

  3. Wait for the AI to generate your letter (usually 10-30 seconds)

  4. Find the generated letter in candidates/example_user/2_applications-mds/YYYY-MM-DD/

Create the PDF

Convert the generated Markdown letter to a professional PDF:

# From the repository root
cd 3_latex

# Copy the generated letter to the content folder
cp ../candidates/example_user/2_applications-mds/2025-01-15/*.md src/content/

# Convert Markdown to LaTeX and compile PDF
make md2pdf

# Find your PDF in output/
ls output/
open output/  # macOS: opens the folder

How It Works

The system uses a structured pipeline to generate professional application letters:

┌─────────────────┐     ┌─────────────────┐     ┌─────────────────┐
│   Job Posting   │     │   Competence    │     │    AI Prompt    │
│   (Markdown)    │     │    Profile      │     │   (GPT rules)   │
└────────┬────────┘     └────────┬────────┘     └────────┬────────┘
         │                       │                       │
         └───────────────────────┼───────────────────────┘
                                 │
                                 ▼
                    ┌────────────────────────┐
                    │      n8n Workflow      │
                    │  (Merges & sends to    │
                    │       OpenAI API)      │
                    └────────────┬───────────┘
                                 │
                                 ▼
                    ┌────────────────────────┐
                    │   Generated Letter     │
                    │     (Markdown with     │
                    │    LaTeX formatting)   │
                    └────────────┬───────────┘
                                 │
                                 ▼
                    ┌────────────────────────┐
                    │   parse_md_to_tex.py   │
                    │  (Converts to LaTeX)   │
                    └────────────┬───────────┘
                                 │
                                 ▼
                    ┌────────────────────────┐
                    │     XeLaTeX Compile    │
                    │   (Creates final PDF)  │
                    └────────────────────────┘

Key Components

  1. AI Prompt (0_queries&prompts/prompts/3_last_update_ATS_optimized_prompt.md)

    • Instructs GPT to write letters with specific sections
    • Uses HTML comment labels: <!-- PARAGRAPH_1_INTRODUCTION -->
    • Formats content in \lettercontent{...} blocks for LaTeX

  2. n8n Workflow (4_n8n/application-letter-agent.json)

    • Reads your job posting and competence profile
    • Merges them with the AI prompt
    • Sends to OpenAI API
    • Saves the generated letter

  3. Markdown Parser (3_latex/parse_md_to_tex.py)

    • Reads section labels from generated markdown
    • Maps them to LaTeX template placeholders
    • Auto-fills your personal info from user_info.yml

  4. LaTeX Templates (3_latex/src/templates/)

    • Professional document layouts
    • Placeholders like {{PARAGRAPH_1_INTRODUCTION}} get replaced with content

File Reference

Files You Need to Create/Edit

File Purpose How to Create
3_latex/user_info.yml Your contact information Copy from user_info.example.yml, edit with your data
candidates/yourname/1_profile/competence-profile.md Your professional profile Copy example folder, customize the profile
candidates/yourname/0_inbox-jobs/DATE/*.md Job postings Copy from 0_template_jobs.md, paste job description
candidates/yourname/.latest_job.json Current job folder path Edit to point to today's job folder

Files That Are Generated (Don't Edit)

File Purpose
candidates/*/2_applications-mds/*.md AI-generated letters
3_latex/src/applications/*.tex Converted LaTeX files
3_latex/output/*.pdf Final PDF documents

Configuration Files

File Purpose
4_n8n/docker-compose.yml n8n Docker settings (edit paths for your system)
4_n8n/application-letter-agent.json n8n workflow (import into n8n)
.gitignore Prevents committing personal data

Customization Guide

Adding a New Candidate

To set up the system for another person (family member, client):

  1. Create their folder structure:

    mkdir -p candidates/newperson/{0_inbox-jobs,1_profile,2_applications-mds,3_job_analysis}
cp candidates/example_user/0_inbox-jobs/0_template_jobs.md candidates/newperson/0_inbox-jobs/
cp candidates/example_user/1_profile/competence-profile.md candidates/newperson/1_profile/
echo '{"jobs": []}' > candidates/newperson/.latest_job.json

  2. Edit their competence profile with their information

  3. Add their info to 3_latex/user_info.yml:

    - id: newperson
  full_name: New Person Name
  # ... rest of fields

  4. Duplicate the n8n workflow and update the file paths to point to the new folder

Modifying the AI Prompt

The AI prompt controls how letters are written. To customize:

  1. Open 0_queries&prompts/prompts/3_last_update_ATS_optimized_prompt.md
  2. Study the structure:
    • Language detection rules
    • Section definitions (introduction, technical, experience, strategic fit, closing)
    • Formatting requirements
  3. Make changes and test with a few job postings
  4. Update the prompt in the n8n workflow node as well

Creating Custom LaTeX Templates

  1. Copy an existing template from 3_latex/src/templates/
  2. Modify the design (requires LaTeX knowledge)
  3. Update parse_md_to_tex.py to use your template

Daily Workflow

  1. Find a job posting - LinkedIn, company websites, job boards
  2. Save the job - Create a markdown file in your 0_inbox-jobs/DATE/ folder
  3. Update .latest_job.json - Point to today's job folder
  4. Run the workflow - Execute in n8n, wait for generation
  5. Review the letter - Check the generated markdown, edit if needed
  6. Generate PDF - Run make md2pdf in the 3_latex/ folder
  7. Send application - Use the PDF, track in your job search spreadsheet

Troubleshooting

n8n Can't Find Files

Problem: Workflow fails with "file not found" errors

Solution:

  1. Check that Docker volume paths in docker-compose.yml are correct
  2. Verify paths in workflow nodes match your folder structure
  3. Ensure the repository is mounted at /workspace/jobs inside the container
  4. Run docker compose down && docker compose up -d to restart

LaTeX Compilation Errors

Problem: make md2pdf fails with errors

Solution:

  1. Check 3_latex/build/*.log for detailed error messages
  2. Common issues:
    • Missing LaTeX packages: Install with tlmgr install packagename
    • Special characters: Escape %, $, &, # in your content
    • Missing fonts: Install required fonts or change the template

AI Generates Poor Quality Letters

Problem: Letters don't match the job or sound generic

Solution:

  1. Improve your competence profile with more specific details and metrics
  2. Ensure the job posting is complete (not just a title)
  3. Check that the AI prompt matches your industry/role
  4. Consider using GPT-4 instead of GPT-3.5 for better quality

Docker Issues

Problem: Docker compose fails to start

Solution:

  1. Ensure Docker Desktop is running
  2. Check port 5678 is not in use: lsof -i :5678 (macOS/Linux)
  3. Try docker compose down && docker compose up -d
  4. Check Docker logs: docker compose logs

OpenAI API Errors

Problem: "API key invalid" or "quota exceeded"

Solution:

  1. Verify your API key at https://platform.openai.com/api-keys
  2. Check your usage limits at https://platform.openai.com/usage
  3. Update the credential in n8n if you regenerated the key

Contributing

Contributions are welcome! Please:

  1. Fork the repository
  2. Create a feature branch
  3. Make your changes
  4. Submit a pull request

Important: Never commit personal data (real names, emails, phone numbers, job applications).

License

MIT License - See LICENSE file for details.

Acknowledgments

  • Built with n8n workflow automation
  • Powered by OpenAI GPT models
  • PDF generation via LaTeX/XeLaTeX

About

No description or website provided.

Topics

application-letters resume-builder n8n motivation-letter llm

Resources

Readme

License

MIT license
Activity
Custom properties

Stars

1 star

Watchers

0 watching

Forks

0 forks
Report repository

Releases

No releases published

Packages

No packages published

Languages