ChatGPT runs its own search index family, Labrador, and is A/B testing it against scraped results
VeryWellVersed · x · 2026-09-06
Peec AI's analysis upends the SEO assumption that ChatGPT only retrieves via Bing and Google: OpenAI has spent years building its own retrieval stack, internally called Labrador — not one index but a family of vertical indexes covering web, PDF, YouTube, news, arXiv, Wikipedia, local, finance, legal, medical, shopping, and images. Between May 21 and July 21, 2026, a resultsource field in ChatGPT server-side events showed only four values: Labrador, Bright, Oxylabs, and SERP, the latter three being external scraping providers. Google antitrust testimony and job postings confirm a deliberate years-long build; ChatGPT is now A/B testing its own index against scraped results, most visibly in shopping. It also caches pages instead of fetching live, and pulls from at least eight other providers including Yelp, TripAdvisor, and Microsoft's Web IQ. Key takeaway: a ChatGPT strategy stopping at Bing misses most of the system.
More from Companies & People
- Senior researcher rebuts 'one-person pretraining' fear: your fundamentals matter more, not less — RexDouglass · 2026-10-23
- Anthropic proactively issues batch refunds and credit notes to a user — TechPreacher · 2026-10-23
- Manycore Aholo Global 3D AI Hackathon opens with $47,800 prize pool — Scobleizer · 2026-10-23
- Thousands built claw machines across Europe as the tour wrapped at GTC Berlin — lifebypixels · 2026-10-23
- Instinct personal agent growing 10% daily, yet resort board had never heard of it — johncoogan · 2026-10-23
- WSJ: OpenAI fired 3 safety researchers over disappearing Signal chats on model Astra — rohanpaul_ai · 2026-10-23