convez.ai
SkillAutomaçãoIntermediário
palych65

Conversor em Lote de PDFs para Markdown

palych65/pdf-folder-to-markdown

Converte lotes de PDFs em Markdown com um comando para Claude

Ver no GitHub
InicialPluginScore · de 0 a 100

Skill para Claude que converte automaticamente todos os PDFs de uma árvore de pastas em Markdown, com script opcional para limpeza posterior. Usa extração de texto simples via PyMuPDF, sem OCR ou reconstrução de layout, ideal para documentação em massa e organização de arquivos.

Instalação

$npx convez-ai palych65/pdf-folder-to-markdown

1 comando no terminal: detecta o formato do repositório (marketplace ou plugin avulso) e instala via Claude Code. Requer Node 18+ e o Claude Code instalado.

Métricas

Estrelas
0
Forks
0
Acompanhando
0
Saúde
42%
Último commit
há 5 meses
Versão
—
Issues abertas
0
Contribuidores
1

Casos de uso

  • Converter acervos de documentos PDF em arquivos Markdown para versionamento
  • Organizar pastas de PDFs agrupadas por cliente/projeto em Markdown único por pasta
  • Automatizar migração de documentação de PDF para formato Markdown para repositórios

Veredito

Solução prática e sem dependências externas pesadas para conversão em lote de PDFs. Força no caso de uso específico (um Markdown por pasta) e scripts reutilizáveis; limitação clara na extração de texto simples sem OCR ou layout.

README

A tradução para o português deste README ainda está na fila.

pdf-folder-to-markdown

A Claude Agent Skill that bulk-converts every PDF in a folder tree to Markdown, with an optional companion step to tidy up the originals afterwards.

Give Claude a folder full of PDFs and ask it to convert them — the skill writes two small, self-contained Python scripts into that folder, runs the converter, and hands you one Markdown file per subfolder. It's deliberately boring under the hood: PyMuPDF's plain-text extractor, no OCR, no layout reconstruction, no cloud calls.

How a skill works

A skill is just a folder with a SKILL.md at its root. The YAML frontmatter gives it a name and a description, and Claude reads that description to decide when the skill is relevant. When you ask something that matches — "convert all the PDFs in this folder to markdown", "turn these PDFs into .md", "now delete the originals" — Claude loads SKILL.md and follows the steps inside it.

This repo is the skill: drop the whole folder in and you're done.

Install

Place this folder in your Claude skills directory so the agent can discover it — the folder name should match the skill name (pdf-folder-to-markdown):

git clone https://github.com/palych65/pdf-folder-to-markdown.git ~/.claude/skills/pdf-folder-to-markdown

Skill discovery locations differ slightly between clients (Claude Code, the desktop app, plugins/marketplaces), so check the Agent Skills docs for where yours looks. The only hard requirement is that SKILL.md sits at the root of the skill folder, which it does here.

Once it's in place, just ask Claude in natural language:

Convert all the PDFs in this folder to markdown.

Claude handles the rest — installing PyMuPDF, running the converter, and offering the cleanup once it's finished.

What it does

The skill drops two scripts into your target folder:

  • templates/convert_pdfs_to_md.py walks the folder tree and converts every *.pdf to Markdown. The output is named after the containing folder, not the PDF — so a folder of PDFs collapses into a single .md, which is handy when each subfolder is one logical bucket (a client, a matter, a borough, whatever you're organising by).
  • templates/delete_non_markdown.py previews and then removes every non-.md file in the tree. It runs as a dry-run by default and only deletes when you pass --apply, so you can't wipe your originals by accident.

Repository layout

pdf-folder-to-markdown/
├── SKILL.md                      # the skill definition Claude reads and follows
├── README.md                     # this file
├── LICENSE                       # MIT
├── .gitignore
└── templates/
    ├── convert_pdfs_to_md.py     # PDF → Markdown converter
    └── delete_non_markdown.py    # optional cleanup (dry-run by default)

Requirements

  • Python 3.8+
  • PyMuPDF for the conversion step (pip install pymupdf). The cleanup script uses only the standard library.

Running the scripts directly

You don't need an agent — the scripts stand on their own. Copy them into the folder you want to process and run them from there.

python3 convert_pdfs_to_md.py                 # convert everything
python3 convert_pdfs_to_md.py --force         # overwrite existing .md files
python3 convert_pdfs_to_md.py --only <name>   # process just one folder by name

Then, if you want to clear out the originals:

python3 delete_non_markdown.py                # dry run — shows what w
Ver o original no GitHub

Bom para

Relacionados