resumejson_cli

A CLI tool to convert resumes between PDF / JSON / Markdown and run an AI-powered ATS pipeline (score → questions → answers → review → optimize).

The canonical format is a ResumeJson object validated by a Zod schema. Every command reads or writes that JSON, so any step in the pipeline can be inspected, edited, or fed into another tool.

Key Features

• Format conversion — extract a structured ResumeJson from a PDF (using a vision model over rendered pages) or Markdown, and render it back to PDF (Mustache + Puppeteer) or Markdown.
• AI-powered ATS pipeline — score, review, generate screening questions, auto-answer them, and rewrite the resume to improve ATS performance.
• Composable JSON I/O — every command takes -i <input> and -o <output>; pass - for stdin/stdout to chain commands.
• Cached PDF rendering — the PDF→images step is memoised in a local SQLite cache, so re-running from_pdf on the same PDF is cheap.
• Bundled Claude skill — ships a SKILL.md that maps the CLI to a skill installable into a Claude agent folder.

Requirements

• Node.js 20+
• OPENAI_API_KEY environment variable — every AI-backed command uses the OpenAI AI SDK (default model gpt-4.1).

Installation

npm install -g resumejson_cli

Or install just the bundled SKILL.md into an agent folder so an AI agent can drive the CLI:

npx resumejson_cli install <skill-folder>

This copies SKILL.md to <skill-folder>/skills/resume-json/SKILL.md.

Quick Start

export OPENAI_API_KEY=sk-...

# PDF -> structured resume JSON
npx resumejson_cli from_pdf -i ./resume.pdf -o ./resume.json

# Score how ATS-friendly the resume is
npx resumejson_cli ats_score -i ./resume.json -o ./resume.ats_score.json

# Render the JSON back to PDF
npx resumejson_cli to_pdf -i ./resume.json -o ./resume.rendered.pdf

Commands

All commands take -i <input> and -o <output>. AI-backed commands also print a human-readable summary to stdout in addition to writing the JSON output file. - is accepted as <path> for stdin/stdout.

| Command | Purpose | Required flags | |---|---|---| | from_pdf | Extract resume JSON from a PDF (vision model) | -i, -o | | to_pdf | Render resume JSON to a PDF | -i, -o | | from_markdown | Extract resume JSON from a Markdown file | -i, -o | | to_markdown | Render resume JSON to Markdown | -i, -o | | ats_score | Numerical ATS readiness score (use as before/after gauge) | -i, -o | | ats_review | Qualitative ATS review (strengths, weaknesses, suggestions); feeds ats_optimize | -i, -o | | ats_question | Generate the ATS-style screening questions a recruiter would ask | -i, -o | | ats_answering | Auto-answer ATS questions using the resume as context | -i, -q, -o | | ats_answered | Fold answered questions back into a richer resume JSON | -i, -q, -o | | ats_optimize | Rewrite the resume guided by an ATS review | -i, -r, -o | | install [skill-folder] | Install bundled skills into <skill-folder>/skills/ (default: .) | — |

Workflow: full ATS pipeline

The recommended end-to-end pipeline, mirroring the full_pipeline script in package.json:

# 1. ingest
npx resumejson_cli from_pdf      -i resume.pdf                                    -o resume.json

# 2. baseline score (optional, for before/after comparison)
npx resumejson_cli ats_score     -i resume.json                                   -o resume.ats_score.json

# 3. enrich with answered ATS questions
npx resumejson_cli ats_question  -i resume.json                                   -o questions.unanswered.json
npx resumejson_cli ats_answering -i resume.json -q questions.unanswered.json      -o questions.answered.json
npx resumejson_cli ats_answered  -i resume.json -q questions.answered.json        -o resume.answered.json

# 4. review and optimize
npx resumejson_cli ats_review    -i resume.answered.json                          -o resume.ats_review.json
npx resumejson_cli ats_optimize  -i resume.answered.json -r resume.ats_review.json -o resume.optimized.json

# 5. render the optimized result
npx resumejson_cli to_pdf        -i resume.optimized.json -o resume.optimized.pdf
npx resumejson_cli to_markdown   -i resume.optimized.json -o resume.optimized.md

# 6. (optional) confirm the score improved
npx resumejson_cli ats_score     -i resume.optimized.json -o resume.optimized.ats_score.json

Scripts

npm run build              # tsc compile
npm run typecheck          # tsc --noEmit
npm run test               # node --test on tests/**/*.test.ts
npm run dev                # run the CLI from source via tsx
npm run full_pipeline      # end-to-end pipeline against the sample PDF
npm run clean:outputs      # remove generated files in ./outputs

Output & Errors

• Conversion outputs (to_pdf, to_markdown) write the rendered file; all other commands write structured JSON.
• AI-backed commands also print a human-readable pretty-printed summary to stdout.
• Use - for either -i or -o to read from stdin or write to stdout.
• Input JSON is validated against ResumeJsonSchema; mismatches surface as Zod validation errors.
• Missing OPENAI_API_KEY causes every AI command to fail — export it before retrying.

License

ISC