State of GEO: how ready is the web for AI search?
Live aggregates from real audits run with the open-source GEO Optimizer engine — the same 100-point, 8-category rubric for every site. No domains, no identities: the dataset stores only salted hashes and per-category signals, and this page shows aggregates only.
Based on — audited domains (— audits) · live dataset, updated as audits run · last refresh —
Average GEO score
—
out of 100
Median score
—
half the web does worse
Ship an llms.txt
—
of audited domains
Adoption of AI-readiness signals
Where sites lose points
Average score per category vs. its ceiling — the widest gaps are the easiest wins.
Score bands
Methodology
- One row per domain — the most recent audit per domain; re-audits don't inflate the cohort.
- Scheduled monitoring excluded — only user-initiated audits (web, CLI, API, tools) count.
- Privacy-first — domains are stored as salted HMAC hashes; this endpoint exposes aggregates only.
- Same rubric for everyone — the open-source engine's 100-point score across 8 categories (robots, llms.txt, schema, meta, content, brand & entity, signals, AI discovery).
The dataset grows with every audit. Run one and you're (anonymously) in it.
Where does your site stand?
The averages above are the baseline to beat. A free audit gives you your score in seconds — then the gap analysis tells you which category recovers the most points first.