source ~/.local/opencode-venv/bin/activate

Need to process a PDF?
│
├─ Extract everything from a PDF?
│  └─ Dispatch the `pdf-extractor` agent.
│     It converts all pages to images, reads each one, and stores
│     per-page results in extracted_data/<pdf_name>/page_NNN/content.md
│     along with images and a summary.
│
├─ Quick text from a normal PDF?
│  ├─ Simple: pdftotext input.pdf output.txt
│  └─ With page numbers: python3 scripts/extract_text.py input.pdf
│
├─ Extract tables?
│  └─ python3 scripts/extract_tables.py input.pdf -o ./tables/
│
├─ Fill a form?
│  ├─ Check fields: python3 scripts/check_forms.py input.pdf
│  └─ Fill: python3 scripts/fill_form.py input.pdf fields.json output.pdf
│
├─ Read a scanned PDF (quick)?
│  ├─ Convert: python3 scripts/to_images.py scanned.pdf -o ./pages/ --dpi 200
│  └─ Read the images and extract what you need
│
├─ Merge/Split/Rotate?
│  └─ See pypdf examples below
│
└─ Large file (>50MB)?
   ├─ Quick text: pdftotext (fastest)
   └─ Page ranges: pdftotext -f N -l M input.pdf

pdftotext document.pdf - | head -20

# All pages at 200 DPI (sufficient for most documents)
python3 scripts/to_images.py document.pdf -o ./pdf-pages/ --dpi 200

# Higher quality for handwritten or low-quality scans
python3 scripts/to_images.py document.pdf -o ./pdf-pages/ --dpi 300

rm -rf ./pdf-pages/

# Simple extraction
pdftotext input.pdf output.txt

# Single page
pdftotext -f 3 -l 3 input.pdf page3.txt

# With layout preservation
pdftotext -layout input.pdf output.txt

# Page range
pdftotext -f 5 -l 10 input.pdf pages_5_10.txt

python3 scripts/extract_text.py input.pdf

pdfinfo input.pdf

python3 scripts/extract_tables.py input.pdf -o ./tables/

import pdfplumber
import csv

with pdfplumber.open("input.pdf") as pdf:
    for i, page in enumerate(pdf.pages):
        tables = page.extract_tables()
        for j, table in enumerate(tables):
            with open(f"page{i+1}_table{j+1}.csv", "w", newline="") as f:
                writer = csv.writer(f)
                for row in table:
                    writer.writerow([cell or "" for cell in row])

python3 scripts/check_forms.py input.pdf

python3 scripts/fill_form.py input.pdf fields.json output.pdf

{
  "field_name": "John Smith",
  "email": "john@example.com",
  "date": "2026-05-04"
}

from pypdf import PdfReader, PdfWriter

writer = PdfWriter()
for pdf_file in ["file1.pdf", "file2.pdf", "file3.pdf"]:
    reader = PdfReader(pdf_file)
    for page in reader.pages:
        writer.add_page(page)

with open("merged.pdf", "wb") as f:
    writer.write(f)

from pypdf import PdfReader, PdfWriter

reader = PdfReader("input.pdf")
for i, page in enumerate(reader.pages):
    writer = PdfWriter()
    writer.add_page(page)
    with open(f"page_{i+1:03d}.pdf", "wb") as f:
        writer.write(f)

from pypdf import PdfReader, PdfWriter

reader = PdfReader("input.pdf")
writer = PdfWriter()
for page in reader.pages:
    page.rotate(90)  # 90, 180, or 270 degrees
    writer.add_page(page)

with open("rotated.pdf", "wb") as f:
    writer.write(f)

from pypdf import PdfReader, PdfWriter

reader = PdfReader("input.pdf")
writer = PdfWriter()
pages_to_keep = {0, 1, 4, 5}  # 0-indexed
for i, page in enumerate(reader.pages):
    if i in pages_to_keep:
        writer.add_page(page)

with open("trimmed.pdf", "wb") as f:
    writer.write(f)

python3 scripts/metadata.py input.pdf

python3 scripts/to_images.py input.pdf --dpi 200

pdftotext input.pdf - | grep -n "search term"

pdftotext doc1.pdf - > /tmp/doc1.txt
pdftotext doc2.pdf - > /tmp/doc2.txt
diff /tmp/doc1.txt /tmp/doc2.txt

from pypdf import PdfReader

reader = PdfReader("encrypted.pdf")
if reader.is_encrypted:
    reader.decrypt("")  # Try empty password first

try:
    with pdfplumber.open("file.pdf") as pdf:
        # process...
except Exception as e:
    print(f"Error processing PDF: {e}")