document-to-mxcp

To ingest documents into an MXCP server, this skill analyzes Excel and Word files to automatically generate DuckDB-backed models for tabular data or RAG-text pipelines for narrative content, producing reproducible dbt or Python scripts for both queryable databases and semantic search.

The best way to convert tabular data for RAG pipelines is to analyze the Excel content and convert it into narrative form, allowing the generation of reproducible RAG-text pipelines that prepare the data for semantic search and retrieval within an MXCP project.

Ingesting documents into MXCP uses DuckDB for structured database models and dbt for reproducible pipeline scripts, deciding automatically between generating DuckDB databases or RAG pipelines based on whether the input content is tabular or narrative.

Yes, you can use Python scripts to automate document ingestion for MXCP projects, relying on standard packages like pandas, python-docx, openpyxl, and duckdb to analyze content and generate reproducible extraction pipelines.

Yes, you need the mxcp-expert skill for project setup, as the document ingestion process depends on it to establish the MXCP server environment before generating the necessary database or RAG components.

You should use a DuckDB database model for tabular content and a RAG pipeline for narrative content, with the ingestion process automatically analyzing the Word document to decide the appropriate extraction path and generating the corresponding dbt or Python scripts.