LarpMyResume

Rule-based, offline resume auditor for ATS compatibility and keyword matching.

No API keys. No LLMs. No external services. Runs in seconds.

LarpMyResume audits your resume PDF for ATS parse reliability, scores your bullet points against an evidence-graded rubric, and compares your skills against any job description entirely offline, entirely deterministically.

Origin story: This was originally going to be a Claude Code skill that called an LLM on every resume scan. Then I looked at my token usage and decided I'm not that rich. Everything is deterministic rules and regex now which honestly works better anyway.

---

The Research Behind It

The scoring rules, thresholds, and flags in this tool are grounded in peer-reviewed research on recruiter behavior, ATS parsing, and candidate screening.

See RESEARCH.md for the full evidence base: a synthesis of ATS vendor documentation, eye-tracking studies, hiring outcome research, and legal guidance on Canadian and US hiring practices. If you disagree with a rule, that's where to look — and where to open a PR if you have better evidence.

---

What It Checks

ATS compatibility (scan)

• Multi-column layout detection (column-gap heuristic using character x-positions)
• Non-standard section headings that ATS systems fail to classify
• Tables, text boxes, and graphics that break parser output
• Critical information in headers/footers
• Image-only PDFs (scanned resumes with no extractable text)

Bullet quality (scan)

• Weak opening verbs and passive phrasing
• Missing or vague metrics
• Anti-AI detection flags (generic, filler, or suspiciously polished language)
• Per-bullet scores with specific improvement flags

Keyword matching (match)

• 300+ skill gazetteer across languages, frameworks, infra, and tools
• Three-tier output: strong matches, weak/partial matches, and gaps
• Short-skill disambiguation (distinguishes "Go" from prose, "R" from sentences)
• Works against a URL or a local text file — no internet required for --job-text

---

Install

Python CLI

pip install larp-my-resume

Ruby Report Generator (optional: for HTML reports)

gem install larp-my-resume-report

---

Usage

ATS audit: scan your resume:

larp-my-resume scan resume.pdf

Keyword match against a job posting URL:

larp-my-resume match resume.pdf --job-url "https://jobs.example.com/posting"

Keyword match against a local job description:

larp-my-resume match resume.pdf --job-text job.txt

Full pipeline: JSON output piped to an HTML report:

larp-my-resume match resume.pdf --job-url "https://jobs.example.com/posting" --json \
  | larp-my-resume-report --out report.html

JSON output (for scripting or piping):

larp-my-resume --json scan resume.pdf
larp-my-resume --json match resume.pdf --job-text job.txt

---

How It Works

pdfplumber extracts character-level text from your resume PDF. A column-detection heuristic clusters character x-positions to flag multi-column layouts. Rule-based analyzers run against the extracted text:

• ATS checker: regex patterns and structural heuristics against known ATS failure modes
• Bullet scorer: verb lists, metric patterns, and flag rules derived from the research in RESEARCH.md
• Keyword matcher: a curated skill gazetteer with short-token disambiguation and importance weighting

The Ruby gem (larp-my-resume-report) reads the JSON output and renders a self-contained HTML report with inline CSS — no JavaScript, no external resources.

---

Development

# Python
pip install -e .
pytest

# Ruby gem
cd ruby && bundle install && bundle exec rspec

# Generate test fixture PDFs
python tests/fixtures/generate_fixtures.py

Tests require no network access and no API keys. All test fixtures are programmatically generated.

---

Contributing

MIT licensed - issues and PRs welcome.

Open an issue before submitting large changes: github.com/B2707/LarpMyResume/issues

If you have better evidence for any scoring rule, the right place to start is RESEARCH.md.

---

License

MIT — © 2026 Bader Asadi