pdf-to-text

What it does

Extract text from PDFs, either from a URL or base64-encoded PDF data.

AgentPM manifest

See agent.json in this folder. It declares:

• "kind": "tool", "runtime": { "type": "python", "version": "3.11" }
• "entrypoint": { "command": "python", "args": ["-u", "pdf_to_text/__main__.py"] }
• JSON Schema for inputs and outputs

Quirks

Entrypoint reads JSON from stdin and prints a single JSON object to stdout.

Setup & run

# install deps (uv)
uv sync --directory tools-python/pdf-to-text

# run
uv run --directory tools-python/pdf-to-text \
python -m pdf_to_text <<'JSON'
{"pdf_url": "https://ontheline.trincoll.edu/images/bookdown/sample-local-pdf.pdf"}                     
JSON

Vendor Dependencies before publishing

uv pip install \
  --target tools-python/pdf-to-text/pdf_to_text/_vendor \
  "pypdf>=5.0.0" \
  "httpx>=0.27.0"