Documents Skill
Process PDFs with native vision - text, images, charts, tables.
Features
- •Analyze up to 1000 pages per PDF
- •Extract structured data
- •Summarize and Q&A
- •Transcribe with formatting
Usage
python
from rhea_noir.skills.documents.actions import skill as docs
# Summarize a PDF
result = docs.summarize("path/to/document.pdf")
# Ask questions
result = docs.ask("path/to/document.pdf", "What are the key findings?")
# Extract structured data
result = docs.extract(
"path/to/invoice.pdf",
fields=["total", "date", "items"]
)
# Compare multiple PDFs
result = docs.compare(
["paper1.pdf", "paper2.pdf"],
question="What are the differences?"
)
Limits
- •50MB or 1000 pages per PDF
- •258 tokens per page