pdf-extract

name	pdf-extract
description	Extract text from PDF files or folders of PDFs and save as sidecar .txt files

PDF Extract

Extract text from PDF files — either a single .pdf file or all PDFs recursively under a folder — and save each result as a .txt file alongside the source PDF.

Instructions

When the user asks to extract text from a PDF file or from PDFs in a folder, run the following command:

dotnet run "${CLAUDE_SKILL_DIR}/pdf-extract.cs" <rootFolder|file.pdf>

${CLAUDE_SKILL_DIR} resolves to this skill's directory regardless of where it's installed (project-level .claude/skills/, user-level ~/.claude/skills/, etc.). Replace <rootFolder|file.pdf> with the absolute or relative path to a directory or a single .pdf file.

Behaviour

• Directory mode: recursively finds every .pdf file under the given folder and all subfolders.
• Single-file mode: processes only the specified .pdf file directly.
• For each PDF, creates a sidecar text file named <filename>.pdf.txt in the same directory as the source PDF.
• Skips any PDF that already has a corresponding .txt file (idempotent — safe to re-run).
• Logs one tagged line per file: [OK], [SKIP], or [ERROR].
• Prints a summary at the end: total files found, processed, skipped, and errors.

Output Example

Scanning: /data/documents

[SKIP]  /data/documents/archive/old-report.pdf
[OK]    /data/documents/2025/annual-report.pdf
[ERROR] /data/documents/corrupt.pdf
        PdfDocumentEncryptedException: The document is encrypted and cannot be read without a password.

Done. Total: 3 | Processed: 1 | Skipped: 1 | Errors: 1

When to Use

• Bulk PDF ingestion pipelines where extracted text is needed for indexing or analysis.
• Direct extraction from a single PDF file without scanning a whole directory.
• Pre-processing step before feeding document content to Claude.
• Any workflow requiring plaintext versions of PDF files without manual extraction.

Prerequisites

• .NET 10 SDK installed and dotnet available on the PATH.
• The UglyToad.PdfPig@0.1.13 NuGet package is restored automatically on first run.
• Write access to the directories containing the PDFs (for creating .txt output files).

Limitations

• Does not extract text from scanned/image-only PDFs (no OCR). Those will produce empty or near-empty .txt files.
• Password-protected PDFs will produce an [ERROR] entry and be skipped.
• Very large PDFs may take noticeable time; extraction is sequential and single-threaded.