What is GEO (Generative Engine Optimization)?

GEO is the practice of making your website readable, citable, and trustworthy to AI answer engines like ChatGPT, Gemini, Claude, and Perplexity. Traditional SEO optimizes for a search engine that returns a list of blue links. GEO optimizes for engines that read your content, decide what's true, and synthesize a single answer. Different audience, different rules.

How is GEO different from traditional SEO?

SEO rewards keyword targeting, backlinks, and ranking on a results page. GEO rewards being the source AI models choose to quote. That means structured data, factual density, named-entity clarity, server-rendered content, and authoritative citations matter more than meta keywords. A page can rank #1 on Google and still be invisible inside ChatGPT — they evaluate completely different signals.

Which AI answer engines does IndexGraph optimize for?

The big four: ChatGPT (OpenAI), Perplexity, Google Gemini, and Claude (Anthropic). We also audit for Google AI Overviews, which now appear on a significant share of Google searches. Each engine crawls and synthesizes content differently, so IndexGraph measures readiness against the patterns each one rewards — not a single generic "AI score."

What is the AI Readiness Score and how is it calculated?

It's a single 0–100 number that summarizes how well your site performs across seven dimensions: AI Readability, Content Semantics, Citation Readiness, AI Access, AI Crawlability Intelligence, E-E-A-T Signals, and Entity Knowledge Graph. Each dimension is scored independently against 200+ checkpoints, then combined using a weighted formula. The breakdown tells you exactly where to focus first.

How long does an audit take?

A free basic audit runs in under a minute. A full deep audit — which includes JavaScript rendering, multiple crawler simulations, schema validation, and AI engine probes — usually finishes in 3 to 10 minutes depending on site size. Larger sites with hundreds of pages take a bit longer. You'll see live progress as the pipeline runs, so nothing feels like a black box.

Is there a free plan?

Yes. You can run a free Basic Audit on any public URL without paying — that gives you your AI Readiness score plus the headline findings for each section. The Standard plan (₹799/month) unlocks deeper diagnostics, AI-driven fix recommendations, and report exports. We built the free tier to be genuinely useful, not a teaser.

Can IndexGraph audit any public website?

Yes — any publicly accessible URL works, whether it's WordPress, Shopify, Webflow, Framer, Next.js, or a fully custom stack. The crawler renders JavaScript so single-page apps audit correctly. We can't crawl content behind a login, paywall, or `noindex` directive, since those are explicitly hidden from search engines and AI engines alike.

Does IndexGraph store the data it crawls?

We store the audit results (scores, findings, recommendations) so you can revisit reports and track progress over time. We don't keep full copies of your crawled content beyond what's needed to generate the report. Data is hosted on managed infrastructure with encryption at rest. You can request deletion any time — details are in our Privacy Policy.

Why does my website rank on Google but not in ChatGPT?

Google indexes pages and matches keywords. ChatGPT reads pages, evaluates whether the content is trustworthy, structured, and citation-worthy, and then decides whether to synthesize it into an answer. Modern SaaS sites built in React, Vue, or Webflow often score well on Google but fail AI crawlers because the content lives inside JavaScript that AI bots don't execute. Run an audit — you'll see exactly which signals are missing.

What should I do after I get my audit results?

Start with the P0 fixes — those have the highest weight in your score and usually take a few hours. Then work through P1 items the same week. After each batch of fixes, re-run the audit; you'll see the score move in real time, which tells you what's actually working. The Pro and Premium plans include an AI-generated playbook that prioritizes fixes for you and tracks them over time.

How off-page SEO works in 2026

IndexGraph's off-page hub measures what we can verify, labels what we sample, and says plainly when we have not measured something. This page documents the sources, the limits, and why the citation graph now matters as much as the link graph.

Where the data comes from

Backlinks, referring domains, anchors, domain rank: DataForSEO Backlinks API. Fetched live only when you click Fetch live summary or Full sync. Results are saved to your database — browsing saved data costs nothing after that.
Toxicity: computed by IndexGraph from the fetched sample using our own heuristics (lib/offpage/toxicityEngine.ts) — spam score, link attributes, and pattern flags. This is not a DataForSEO verdict.
Broken Link Builder:IndexGraph's own crawler fetches the prospect page. No third-party index.
Mentions & outreach pipeline: your own records, stored in your account.

What we sample, and what that means

A Full sync fetches up to 500 backlinks, 500 referring domains and 200 anchor texts. Toxicity is therefore measured on a sample, not on your entire profile.

We always show the denominator: "20 toxic in 500 sampled (of 48,000 total backlinks)". If we have not scanned your links, we say Not assessed — we never show a score we have not measured.

A "100%" you have not earned is worse than no number at all.

Toxicity carries the date it was measured. A summary refresh updates your backlink totals but does not re-scan links, so the toxicity figure keeps its original date until the next Full sync.

What we do not measure

We do not claim to see every backlink. Nobody does — every tool samples a proprietary index.
Toxicity is a heuristic, not a Google penalty prediction. Google does not publish one.
Domain rank / link authority is a DataForSEO scale, not Moz DA or Ahrefs DR. Do not compare across tools.
Disavow files are generated for you to submit to Google — IndexGraph does not submit them.

How off-page actually works in 2026

Backlinks decide Google's authority. They do not decide what an LLM says.

An AI answer is assembled from the sources the engine retrieves and trusts: Reddit, Wikipedia/Wikidata, YouTube, review platforms (G2, Clutch, Capterra), "best X" listicles, news, and industry blogs.

A DA-40 listicle that Perplexity cites beats a DA-80 backlink it never reads.

Off-page splits into two jobs:

Link graph (Google): authority, toxicity, anchors, disavow — what the off-page hub measures today.
Citation graph (AI engines): are you present in the sources the models actually read?

Practical consequences we recommend: fix what is measurably broken; get cited where AI looks; keep your entity facts consistent across those sources (LLMs synthesise — contradictions get you dropped).

IndexGraph's off-page hub covers the link graph with measured data and honest labels. AI visibility probes and citation-source tracking live in other parts of the product; we do not claim the hub alone tells you everything an LLM will say.

How we generate AI recommendations

When you click Generate recommendations, Gemini reads only what is already saved in your database — no extra DataForSEO calls. Unmeasured fields are sent as not_measured; the model is instructed not to infer zeros or declare a profile "clean" from missing data.

Results are cached per snapshot. Regenerating consumes one AI recommendation credit (Pro: 5/month, Premium: 10/month). Browsing a saved recommendation is free.

Cost & limits

Live data pulls: Pro 3/month, Premium 10/month (Fetch live summary and Full sync share this budget).
Competitor gap snapshots: Pro 3/month, Premium 10/month (separate budget).
AI recommendations: Pro 5/month, Premium 10/month (separate LLM budget).
Browsing saved data is unlimited and free.

Live crawls cost real money. We cap pulls rather than degrade quality or pass runaway API bills through to subscribers.

Changelog

Last updated 2026-07-12 — initial publication: sample denominators, toxicity measurement dates, separate competitor and AI recommendation budgets, grounded Gemini recommendations, link graph vs citation graph framing.