Skip to content

PDF Summary (pdf)

Goal: extract text from a PDF in your workspace and write a concise summary.

Prereqs

  • A PDF file in your workspace (or download one with the commands below).
  • A configured LLM provider (Ollama is a good default for local testing).

Get a sample PDF (optional)

1
2
mkdir -p samples
curl -L -o samples/dummy.pdf https://www.w3.org/WAI/ER/tests/xhtml/testfiles/resources/pdf/dummy.pdf

1
2
New-Item -ItemType Directory -Force samples | Out-Null
Invoke-WebRequest -Uri "https://www.w3.org/WAI/ER/tests/xhtml/testfiles/resources/pdf/dummy.pdf" -OutFile "samples/dummy.pdf"

Run

1
loopforge agent run --workspace . --prompt "Use the pdf tool to extract text from samples/dummy.pdf (max_pages=10, max_chars=12000). Then write notes/pdf_summary.md with: (1) a 6-bullet summary, (2) key terms, and (3) any missing/garbled parts you notice. Only use the extracted PDF text; do not invent content."

1
loopforge agent run --workspace . --prompt "Use the pdf tool to extract text from samples/dummy.pdf (max_pages=10, max_chars=12000). Then write notes/pdf_summary.md with: (1) a 6-bullet summary, (2) key terms, and (3) any missing/garbled parts you notice. Only use the extracted PDF text; do not invent content."

Optional: select pages

If you only need a subset, pass a 1-indexed page selector like pages="2-3" (or pages="2,4-6").

What to expect

  • notes/pdf_summary.md