EvidenTrail

EvidenTrail is a web app (with optional desktop agent) that continuously captures and structures regulatory evidence for AI-enabled medical device software. It ingests artifacts you already produce—Git commits, pull requests, CI/CD logs, model training runs, dataset lineage, risk files, and test results—and maps them to FDA QMSR/21 CFR 820 and ISO 13485/IEC 62304/ISO 14971 evidence expectations. The AI layer drafts “evidence packets” (traceability snippets, change impact summaries, verification summaries) and flags gaps like missing approvals, unlinked requirements, or undocumented model updates. It does not promise compliance; it produces organized, reviewable evidence with clear provenance and immutable timestamps. Teams can export auditor-friendly bundles (PDF + linked source artifacts) and maintain a living “audit room” that stays current as development continues. Realistically, it saves weeks of scramble before audits, but it won’t replace a quality manager or fix broken processes.