Data-Forge is a Model Context Protocol (MCP) server that transforms any LLM into a powerful Data Science Assistant. It provides a suite of high-performance tools for data loading, cleaning, validation, profiling, feature engineering, and visualization.
Stars
2
Forks
0
Watchers
2
Open Issues
0
Overall repository health assessment
No package.json found
This might not be a Node.js project
13
commits
fix: Improve server stability with crash handling, Numba configuration, headless Matplotlib, stderr logging, and optional tool context.
95756b6View on GitHubdocs: Update git clone URL and clarify environment configuration heading
0036e0cView on GitHubfix: Remove explicit python command and PYTHONPATH from server run instructions in README.
501b86aView on GitHubfeat: Add real-time progress updates and improve 3D visualization for semantic void scanning.
69884cbView on GitHubfeat: document new `scan_semantic_voids` and `run_sql_query` tools, and expose LLM instructions via a new prompt function.
40986b6View on GitHubfeat: enable SQL querying of dataframes via DuckDB with a new tool and verification script, adding related dependencies.
b112c23View on GitHubstyle: format empty DataFrame check for improved readability.
ced7ed4View on GitHubfeat: introduce `_to_pandas_safe` utility to safely convert large DataFrames to Pandas with sampling, preventing OOM errors in various operations.
edc20e4View on GitHubfeat: Add semantic void discovery using TDA for text analysis and integrate new dependencies.
72483d9View on GitHub