InferMap

InferMap is an inference-driven schema mapping engine, exposed as an MCP (Model Context Protocol) server. It enables AI assistants and compatible agents to map messy or inconsistent dataset columns to a known schema, supporting both Python and TypeScript/Next.js environments. InferMap provides 7 built-in scoring methods, domain-specific dictionaries (healthcare, finance, ecommerce), human-readable mapping rationales, and confidence calibration with high accuracy (F1 0.84). It can be used on CSVs, DataFrames, and database tables, and integrates with data tools and pipelines, making it ideal for data engineers, analysts, and anyone needing robust, explainable schema alignment in their data workflows.