Job Application Automation Stack

Stack Components

| Tool / Service | Type | Role | Verdict |

|---|---|---|---|

| [HeadlessX](https://github.com/saifyxpro/HeadlessX) ★ 958 | Dev Tool | Scrapes job postings into clean markdown | [SOLID](https://repoverifier.dev/reviews/saifyxpro-headlessx) |

| [n8n](https://github.com/n8n-io/n8n) ★ 52.0k | Dev Tool | Orchestrates the entire pipeline end to end | [SOLID](https://repoverifier.dev/reviews/n8n-io-n8n) |

| pdfplumber | Service | Extracts text from your resume PDF | |

| OpenAI GPT-4o-mini | Service | Tailors resume and cover letter honestly per role | |

| weasyprint | Service | Renders tailored documents as PDFs | |

| Resend | Service | Delivers finished PDFs by email | |

*Established services are not reviewed under RepoVerifier Standard v1.0. Used and recommended based on personal experience.*

Proof it works

Built with this exact stack: [github.com/devopsballog25-pixel/apply-pilot](https://github.com/devopsballog25-pixel/apply-pilot)

AI-Executable Build Brief

You are helping a developer set up apply-pilot — an automated job application pipeline that takes a job posting URL and a resume PDF and delivers a tailored resume and cover letter by email. Follow this brief exactly.

What you are building

A self-hosted job application pipeline with:

• HeadlessX on your host machine — scrapes any job posting URL into clean markdown
• resume-parser Docker service — extracts text from your uploaded resume PDF using pdfplumber
• pdf-renderer Docker service — renders HTML documents to PDF using weasyprint
• n8n Docker service — orchestrates the full workflow: form → scrape → parse → LLM → render → email
• OpenAI GPT-4o-mini — rewrites your resume and generates a cover letter for the specific role, using only your real experience
• Resend — delivers both PDFs to your inbox

Prerequisites (install before starting)

1. Docker and Docker Compose

2. Node.js (for HeadlessX CLI)

3. OpenAI API key — get one at https://platform.openai.com/account/api-keys

4. Resend account with a verified sending domain — https://resend.com

Step 1 — Clone the repo

git clone https://github.com/devopsballog25-pixel/apply-pilot
cd apply-pilot

Step 2 — Install and start HeadlessX

HeadlessX runs on your host machine, not inside Docker. Open a dedicated terminal and keep it running throughout.

npm install -g headlessx
headlessx init
# Choose: self-host
# Start with Docker: Yes
headlessx start
# API server now running at http://127.0.0.1:38473
# Dashboard at http://localhost:34872

The API key lives at ~/.headlessx/repo/infra/docker/.env under DASHBOARD_INTERNAL_API_KEY — the CLI does not surface it directly.

Step 3 — Configure environment

cp .env.example .env

Edit .env and set:

OPENAI_API_KEY=your-key-here

Step 4 — Start Docker services

docker-compose up -d

This starts three services:

• resume-parser on port 5001 — pdfplumber-based PDF text extractor
• pdf-renderer on port 5002 — weasyprint-based HTML to PDF renderer
• n8n on port 5678 — workflow orchestrator

Verify all three are running:

docker-compose ps
# All three should show status: Up
curl http://localhost:5001/health
curl http://localhost:5002/health

Step 5 — Import the n8n workflow

1. Open n8n at http://localhost:5678

2. Click Workflows → Import from File

3. Select n8n/workflows/apply-pilot-workflow.json

Step 6 — Edit your contact details

The imported workflow has placeholder credentials. You must update them before running:

1. Open the imported workflow in n8n

2. Find the Code node (the one that builds the OpenAI request body)

3. Locate these two lines:

const candidateEmail = "[email protected]";
const candidatePhone = "(555) 123-4567";

4. Replace with your real email and phone number

5. Click Save on the node

6. Click Save on the workflow (top right)

No restart needed — n8n applies changes immediately.

Step 7 — Configure Resend credential in n8n

1. In n8n, go to Settings → Credentials

2. Add a new SMTP credential:

- Host: smtp.resend.com

- Port: 465

- Username: resend

- Password: your Resend API key

3. In the Send Email node, set From to [email protected] (your verified Resend domain)

Step 8 — Run it

Open the apply-pilot form at http://localhost:5678, paste a job posting URL, upload your resume PDF, enter your email, and submit.

Check your inbox — you will receive two PDFs within ~60 seconds:

• tailored_resume.pdf — your resume rewritten for that specific role
• cover_letter.pdf — a cover letter tailored to that company and position

Verify it works

• [ ] Workflow imported without errors in the n8n UI
• [ ] Code node shows your real email and phone after editing
• [ ] All three Docker services show Up in docker-compose ps
• [ ] HeadlessX terminal shows activity when form is submitted
• [ ] Email received with both PDFs attached

Environment variables

OPENAI_API_KEY=your-openai-api-key

Architecture

Form (Job URL + Resume PDF + Email)
  ├── HeadlessX (host) — scrapes job posting → clean markdown
  └── resume-parser (Docker :5001) — extracts resume text via pdfplumber
         ↓ (both merged in n8n)
  GPT-4o-mini — tailors resume + cover letter (honest rewrite only)
         ↓
  pdf-renderer (Docker :5002) — renders both documents as PDFs via weasyprint
         ↓
  Resend — emails tailored_resume.pdf + cover_letter.pdf to candidate

Known gotchas

• HeadlessX must be running on the host before submitting the form — n8n reaches it via http://host.docker.internal:38473, not localhost
• The HeadlessX API key is at ~/.headlessx/repo/infra/docker/.env under DASHBOARD_INTERNAL_API_KEY — the CLI does not surface it directly
• pdfplumber may miss resume content in headers or sidebars — inject your real contact details directly into the Code node rather than relying on PDF extraction
• weasyprint must be pinned to version 52.0 — v60.0 has a pydyf incompatibility that breaks PDF rendering
• Do not re-export your live n8n workflow back to n8n/workflows/apply-pilot-workflow.json — that file is the public template with placeholder credentials; keep your live workflow separate

Proof this works

apply-pilot was built and tested using exactly this stack. Full pipeline confirmed working end-to-end — job URL in, two tailored PDFs delivered by email within 60 seconds. Live at https://github.com/devopsballog25-pixel/apply-pilot