Table Extraction Intelligence
Extracts structured tabular data from PDFs, spreadsheets, and scanned documents using advanced layout-aware vision models.
The Table Extraction Intelligence skill empowers Claude Code to accurately parse and reconstruct tabular data from complex sources including PDFs, scans, and spreadsheets (.xlsx, .ods, .csv). Leveraging Xberg's layout-aware detection powered by RT-DETR v2 alongside specialized table reconstruction backends like TATR and SLANet, it converts intricate financial statements, invoices, and scientific grids into clean Markdown tables or queryable JSON cell structures with exact bounding coordinates.
주요 기능
01Direct spreadsheet parsing for XLSX, ODS, and CSV formats without OCR overhead
02Layout-aware table detection powered by RT-DETR v2 for high precision
03Configurable reconstruction backends including TATR and SLANet for bordered or borderless grids
04Dual output modes providing inline Markdown tables or structured JSON cell arrays
05Programmatic integration patterns for Python and Node.js document intelligence pipelines
069,469 GitHub stars
사용 사례
01Parsing financial statements and annual balance sheets into structured JSON data
02Ingesting research papers and dense scientific tables directly into LLM context windows as Markdown
03Automating invoice and bank statement processing pipelines with boundary-accurate cell coordinates