Editorial Methodology

Last reviewed 16 June 2026 · maintained by the VRTubbies editorial team.

This page documents how a scene gets indexed on VRTubbies, what standards we apply when writing the on-page copy, and where AI assistance enters the pipeline. We publish this so that you can decide whether to trust the catalogue and so that Google's quality raters can see we treat the site as edited content rather than scaled output.

1. Sourcing

Scenes are sourced from official studio feeds and partner data drops. We do not scrape competitor catalogues. Studios we list have a publicly visible business presence, a 2257 record-keeping statement of their own, and at least five released VR scenes.

2. Scoring rubric

Each studio is scored on five fixed axes. Each axis is rated 1–10 and the composite score determines whether a studio gets a featured page, a basic listing, or no page at all.

3. AI-assistance disclosure

We use large language models to draft on-page summaries from the structured catalogue data (title, runtime, models, categories, format). The model never invents facts about a scene that are not already in the structured data. Drafts are reviewed by an editor before publication; pages that do not pass review remain noindex and are excluded from the sitemap.

Every page that uses AI-assisted copy carries a visible “Editorial note” aside. We do not bury this disclosure in the terms of service.

4. Indexing rules

5. Editor review SLA

New scenes are reviewed by an editor within 7 days of ingest. Studios in our top-25 by composite score are re-audited quarterly. We sample 100 random scenes per quarter and re-score them against this rubric internally to catch model drift.

6. Corrections and takedowns

If you spot an inaccuracy on a scene page, email [email protected] with the URL and we will correct it within 72 hours. For copyright or 2257 issues see our DMCA page.