P2Predict
This parametric pricing tool for procurement and sales learns from a company's purchase history which factors actually drive a part's price (supplier, material, spec, region), and lets an AI agent ask whether a quote is fair, returning a price estimate with a confidence range and a spec-by-spec breakdown. Use cases include data-driven negotiation targets, engineering value workshops (estimating BOM savings from spec changes), and sales bid preparation. It runs as a CLI, ships an MCP server so it plugs into Claude, GPT, or other agent harnesses, and works 100% locally and privately; it is free for internal company use and source-available. The author also validated its math on public TDC ADMET pharma benchmarks (placing 7th of 24 on one board) and used it to predict rent savings when moving apartments.