Can this skill handle images or charts within documents?

Yes, it extracts images using python-docx or openpyxl, analyzes them with vision capabilities to describe the data or trends, and stores that description as RAG context.

Is the data ingestion process reproducible?

Absolutely. The skill generates dbt models and extraction scripts that are idempotent, meaning they can be re-run from scratch to produce identical results from the source files.

Does it support complex Excel structures like merged cells?

Yes, the skill includes specialized logic to handle merged headers, hierarchical structures, and transposed tables to ensure data integrity during the extraction process.

How does it decide between Database and RAG storage?

It analyzes the content: structured tables with numbers or dates intended for aggregation are sent to DuckDB, while narrative text and descriptive content are optimized for RAG semantic search.

Which file formats does the Document to MXCP skill support?

The skill currently supports Excel files (.xlsx, .xls, .csv) and Word documents (.docx). For older .doc files, conversion to .docx is recommended before ingestion.

Document to MXCP Ingestor

Name: Document to MXCP Ingestor
Author: raw-labs

byraw-labs

•

데이터 과학 및 ML

Automates the intelligent ingestion of Excel and Word documents into MXCP servers for structured querying and semantic search.

The Document to MXCP skill provides a sophisticated, code-first pipeline for transforming static files into actionable data within an MXCP project. It intelligently analyzes file content to determine the optimal storage strategy—classifying data for either DuckDB (structured SQL queries) or RAG (narrative semantic search), or both. By handling complex document features like merged cells, hierarchical headers, and embedded images via vision analysis, it ensures that critical context is preserved. The skill prioritizes reproducibility by generating dbt models and automated scripts, allowing teams to maintain a version-controlled, idempotent data environment.

주요 기능

01Reproducible data pipelines using dbt models and uv-managed virtual environments.

02Intelligent content classification for DuckDB storage or RAG-optimized text extraction.

03Automated image extraction and vision-based analysis for charts, diagrams, and tables.

04Multi-format support for Excel (.xlsx, .csv) and Word (.docx) documents.

05Advanced parsing for complex layouts, including merged cells and transposed tables.

062 GitHub stars

사용 사례

01Automating the extraction of structured data from complex spreadsheets with mixed narrative notes.

02Converting legacy business reports and spreadsheets into a queryable MXCP database.

03Preparing unstructured Word documents for AI-powered semantic search and RAG systems.

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add raw-labs/raw-labs-claude-marketplace document-to-mxcp

For use in Claude.ai and ChatGPT

Download Skill