What it does

This skill provides high-performance Optical Character Recognition (OCR) by leveraging the NVIDIA NeMo Retriever API. It allows your AI agent to "see" and extract text from images and documents with professional-grade accuracy. It handles complex structures like tables, charts, receipts, and even handwriting, returning structured text along with confidence scores and bounding box data.

Why use this skill

Standard LLM vision capabilities can sometimes hallucinate text or struggle with small, dense data like tables or low-quality screenshots. This skill uses a specialized OCR model optimized for precision. It supports batch processing of entire directories, provides confidence metrics to ensure data reliability, and automatically saves output to structured files for further analysis. It is significantly faster and more accurate for data extraction tasks than generic vision prompting.

Supported tools

• NVIDIA NeMo Retriever: State-of-the-art OCR foundation model.
• Python Integration: Built-in handling for Base64 encoding and batch file processing.
• Exporting: Saves results locally in .txt or .md formats for easy developer access.