# Depureco USA AI Website Database Package v5 — Live Validation Readiness

Last updated: 2026-06-10

v5 does not add another broad content dump. It turns v4 into a deployable, auditable website AI package with upload manifests, technical SEO controls, canonical/redirect action buckets, public/private data boundaries, and a content sync playbook.

## What to upload
Upload root files and `/ai/` public docs/db files. Do **not** upload `/internal-review/`.

## Technical SEO recommendation
- Keep `/llms.txt`, `/llms-full.txt`, and `/ai-sitemap.xml` accessible.
- Use `X-Robots-Tag: noindex, follow` for `/ai/*.md` and `/ai/db/*.json` if possible.
- Do not block `/ai/` in robots.txt if you want crawlers and AI systems to discover the support files.
- Validate `/ai-sitemap.xml` after upload and optionally add it to robots.txt as `Sitemap: https://depurecousa.com/ai-sitemap.xml`.

## v5 QA snapshot
- Public files in package: 25
- JSON files checked: 18
- JSON invalid files: 0
- Sitemap URLs: 354
- Sitemap AI-package local missing files: 12
- URL inventory records: 6541
- URL watchlist records: 1917
- Product claim records: 349

## Highest-risk areas to validate manually
1. Old ATEX URLs/copy vs current ORD LOC/explosion-proof U.S. naming.
2. Misspelled product slugs or legacy URLs still referenced by AI files.
3. Duplicate long-tail/application pages that should canonicalize to one strongest page.
4. JSON/Markdown files accidentally getting indexed as normal landing pages.
5. Product claims involving HEPA, NFPA, ORD LOC, Class/Division, UL, or certification language.

## New v5 files
- `/ai/db/deployment-manifest-v5.json`
- `/ai/db/technical-seo-ai-controls-v5.json`
- `/ai/db/canonical-redirect-action-map-v5.json`
- `/ai/db/content-to-database-sync-map-v5.json`
- `/ai/db/public-private-boundary-v5.json`
- `/ai/db/page-update-playbook-v5.json`
- `/ai/db/claim-validation-watchlist-v5.json`
- `/ai/v5-live-validation-readiness.md`

Use the internal QA report only for deployment review, not public upload.