Can AI replace TermsFeed?

Build a closest honest personal substitute for TermsFeed in an empty repository.
Use Next.js 15, TypeScript, PostgreSQL, Drizzle ORM, and a versioned Markdown template engine; do not offer alternative stacks.
The core loop is: collect facts, fill clearly labeled user-supplied policy templates, track versions and approvals, and repeatedly state that the output requires qualified review.
Make the first run work locally with one documented command.
Store all user data locally by default and make export straightforward.
Put secrets in .env, ship .env.example, and never commit credentials.
Create versioned questionnaires, reusable fact fields, conditional sections, templates, and approval states.
Show the source, jurisdiction, owner, reviewed date, and expiry date for every template.
Generate DOCX and PDF drafts with prominent draft and review-required labels.
Keep an immutable audit log of answers, template version, generation time, downloads, and approvals.
Add role-based access, retention, export, encrypted backups, and a deadline calendar.
Display legal-review warnings before generation and again before any final download.
Include clear empty, loading, success, and recoverable error states.
Add input validation, safe filenames, and graceful handling of unavailable APIs.
Write focused tests for the core transformation and one end-to-end happy path.
Create a README with setup, architecture, permissions, data location, and backup steps.
Do not add accounts, billing, telemetry, analytics, or a hosted control plane.
Do not claim to reproduce proprietary data, network liquidity, regulated access, or frontier infrastructure.
Deliberately leave out legal advice or guaranteed compliance.
Deliberately leave out government filings, identity verification, and registered-agent services.
Deliberately leave out automatic consent decisions or deployment without qualified review.
Finish by running the tests and listing the exact commands used.