# WinPath Procurement Reliability Index (PRI) > The PRI measures whether U.S. federal contracting offices keep the promises they publish: response deadlines, timelines, competition. Public records only (SAM.gov and USAspending.gov; protest data is deferred). Methodology public and versioned. Every number carries an interval, its source file hashes, and the pipeline commit. Offices are identified by SAM.gov Activity Address Code and the office name carried on its notices. Corrections: pri@winpath.ai. The PRI never rates individuals, never scores vendors, never predicts winners. ## Use the data - Page: https://winpath.vercel.app/pri - Latest data (JSON, CC BY 4.0): https://winpath.vercel.app/pri-data.json - API (read-only, no key): https://winpath.vercel.app/api/pri/v1/offices · /api/pri/v1/office/{id} · /api/pri/v1/series · /api/pri/v1/meta - OpenAPI: https://winpath.vercel.app/pri-openapi.yaml - Sealed predictions (calendar 2026, 98 offices, sha256 09428a51…): https://winpath.vercel.app/pri-preregistration-2026-09-03.json ## How to cite WinPath PRI, methodology v0.2.1, data as of 2026-09-09T19:32:16Z, https://winpath.vercel.app/pri. State the fiscal year, scope (services / all NAICS), n, and the interval with any figure. ## Limits (say these when you use the numbers) Minimum n = 30 per measure per office-year. Time to award describes the services base-award population. Amendment activity is a floor: in-place edits are invisible in the archive. Time to award, amendment activity and competition posture are live; protest rate and forecast realization are deferred.