When buyers ask AI,
does your brand
make the shortlist?

Three steps. One Monday morning.

This is what you receive.

An illustrative dataset rendered through the exact report code a paying customer gets. Click through — it’s a real HTML page, not a screenshot.

Open the sample report →

Is your site invisible?

Our own blog sat unindexed for months because broken internal links severed the crawl paths — and nothing warned us. The free crawl check walks your site the way Googlebot does and shows you every orphan page, broken internal link, and blocked AI crawler in about a minute. Fixes ship as a copy-paste prompt for your coding agent.

Check my site — freeRead the story behind it

Field notes on AI search.

Working notes on answer engine optimization — how ChatGPT, Perplexity, and Gemini decide which brands to recommend, and what the data actually shows.

• Which AI Customer Support Tool Do AI Engines Recommend? We Asked 5 Engines 75 Times

  We asked ChatGPT, Claude, Gemini, Perplexity and Grok which AI customer support tool to use, 75 times. Intercom Fin was named in 47 of 75 answers. Chatbot questions and agent questions had different winners, and ChatGPT answered chatbot questions with OpenAI's own tools.

  7 min read
• Chatbase Rebranded to AI Agents. AI Search Still Calls It a Chatbot.

  We asked ChatGPT, Claude, Gemini, Perplexity and Grok which AI customer support platform to use, 75 times. Chatbase was named in 20 of 30 answers about chatbots and 8 of 45 about AI agents, where Intercom Fin led. ChatGPT never named it, and pointed to OpenAI's own tools instead.

  6 min read
• The AI-Recommended App Stack, 2026: Every Engine Says Next.js, Supabase, Vercel and Stripe

  We asked ChatGPT, Claude, Gemini, Perplexity and Grok what to build an app with, 75 times. 51 answers named the same four products. Here is the stack each engine would build you, layer by layer, and where they disagree.

  7 min read
• AI Engines Pick Resend for Sending App Email. Their Most-Cited Source Is a Vendor Blog.

  We asked ChatGPT, Claude, Gemini, Perplexity and Grok which email API to use for signup and password reset emails, 60 times. Resend was named in 92% of answers, ChatGPT was the exception, and the most cited source was one email vendor's blog.

  6 min read
• Discovered — Currently Not Indexed: The Silent Cause Nobody Checks

  Our blog sat invisible in Google for months with every post stuck at "Discovered — currently not indexed." The cause wasn't content quality or crawl budget — it was broken internal links severing the crawl paths. Here's how to diagnose it in 60 seconds, and why the same failure hides you from AI crawlers too.

  8 min read
• State of AI Search 2026: We Asked 5 AI Engines to Recommend Software

  We ran 10 buying questions through ChatGPT, Claude, Gemini, Perplexity, and Grok — 3 times each, 150 grounded answers — and counted every brand they named. Every category had a locked shortlist, well-known challengers flickered, and 11 real brands were never named once.

  11 min read

Read all field notes →

Questions worth asking.

What does it cost?

Tracking is $49 a month: your first report today, a fresh run every week across all five engines including Claude, Gemini, and Grok, and a dashboard that shows whether your citation rate is moving. Weekly runs cover up to three brands; past that the schedule is monthly. Cancel any time from the billing portal. A single report is $79 if you only want a snapshot.

What is in a report?

Citation rates per prompt, a gap analysis versus your competitors, and concrete content recommendations to close the gaps we found, delivered by email within fifteen minutes. If we can’t generate it, we refund the report automatically within an hour.

Why not just ask ChatGPT myself?

You can. But AI answers vary run to run, web search results shift, and a single conversation is a sample size of one. We run your prompts multiple times across multiple models with web search, then parse every response for brand mentions. You get the trend, not the anecdote.

Is this exactly what I will see in the ChatGPT or Claude app?

No, and any honest visibility tool should say so. We query the providers’ grounded APIs. Consumer apps can use different model versions, system instructions, rollouts, and personalization. The report is a reproducible directional benchmark, not a promise that every user will receive the same answer.

Why track instead of buying one report?

A single run is a snapshot with sampling noise in it. The question that matters is whether the pages you ship move the number, and that takes a second run to answer. Tracking re-runs the same questions every week and charts the answer; a one-off report is there for a quick look, or use the open-source CLI to run it on your own schedule.

Is the underlying CLI open source?

Yes. The full openllmrank CLI is MIT-licensed on npm. Bring your own OpenAI, Anthropic, Gemini, Perplexity, or xAI key and you can self-host the entire system for the cost of a few API calls. We charge for the hosted version because most marketing leads can’t (and shouldn’t need to) run a CLI.

What if my brand has zero citations?

You’ll get the most useful report of all. We’ll show you exactly which competitors are getting cited, on which prompts, and — when the provider returns source citations — which pages may be contributing. The empty state is the value — you can’t fix what you can’t see.