Multimodal Analysis

Extract text from PDFs with charts, diagrams, and images using vision AI.

Setup

Local (Ollama)

# Install from https://ollama.ai
ollama pull llava

Cloud (OpenAI)

Sys.setenv(OPENAI_API_KEY = "sk-...")

Usage

library(TextAnalysisR)

# Extract PDF with images
result <- extract_pdf_multimodal(
  "document.pdf",
  vision_provider = "ollama"  # or "openai"
)

# Use in analysis
tokens <- prep_texts(result$combined_text)

Smart Extraction

Auto-detects document type:

result <- extract_pdf_smart("paper.pdf", doc_type = "auto")

Provider Comparison

Provider	Cost	Privacy	Accuracy
Ollama	Free	Local	Good
OpenAI	Per use	Cloud	Best