GEO

[sethblack]

It's hard to ignore that GEO is becoming a thing. Google has started using AI Overviews for almost half of their search queries. There's not a lot of good documentation out in the world, so I'm going off of https://searchengineland.com/seo-roadmap-ai-search-449199 for a first run at tackling the SEO + GEO problem.

[Prasanthi-dm]

Totally agree — GEO is becoming a huge factor, especially with Google’s AI Overviews now appearing in nearly half of queries.

This article gives a solid starting point: https://searchengineland.com/seo-roadmap-ai-search-449199

Curious how others are preparing their content and metadata for AI-first search results.

[santimaestri]

Hi, anybody successful when running geo with API?

apparently the model used is no longer availabe

[alexander-k-eliot]

Since #130/#132 already split training vs. retrieval/citation crawler access, that's actually a real chunk of a GEO signal already in the tool. A few more deterministic, no-API-needed checks would cover most of what's currently marketed as "GEO":

• llms.txt presence — the emerging convention for a machine-readable index of the site/business
• schema.org structured data (Organization/Product/FAQ) — what most answer engines actually parse for entity facts
• content readable without JS execution — a lot of crawlers/agents don't render client-side JS at all

@santimaestri's point above about the model no longer being available is worth calling out directly: anything that works by live-prompting an LLM to see if it mentions you is inherently non-reproducible and API-dependent (breaks exactly like that when a model gets deprecated). The three checks above are all static/deterministic against the page itself — same category as what #130/#132 already added, just extended.

Happy to open a PR for any of these if it'd be useful, same shape as the last two.