AI visibility trackers vs AI search analytics: two ways to measure LLM visibility

An AI visibility tracker such as Profound, Peec AI, Otterly or Semrush's AI Visibility Toolkit runs prompts through ChatGPT, Perplexity, Gemini and other assistants every day and scores how often the answers mention and cite your brand, on paid plans from $29 to $495 a month (as of 7 November 2026). AgentFirstStack doesn't do that. Its free AI search analytics, which an AI agent sets up, measures what AI crawlers read on your site, what the visitors assistants send do, and what to change. To know what assistants say about you, use a tracker; the two work well together.

Updated 7 November 2026

Set it up with one prompt

This sets up the site side: AI crawlers and the visitors assistants send. It leaves any tracker you already pay for alone, and ends with the pages your agent can line up against the tracker's citations.

Paste this into Claude Code, Cursor or Codex

Set up AI search analytics for this project with AgentFirstStack. It's free, open-source web analytics built for agents that shows which AI crawlers read the site and which AI assistants send visitors, and you can do the setup yourself: there is no signup form. 1. Read https://agentfirststack.com/llms.txt and https://analytics.agentfirststack.com/llms.txt, in particular its "AI crawlers" and "AI search" sections. 2. If this project's .env already has AGENTFIRSTSTACK_API_KEY, use it and read AGENTFIRSTSTACK.md. Otherwise get a key: curl -X POST https://agentfirststack.com/v1/signup, and save api_key in .env as AGENTFIRSTSTACK_API_KEY and in the production secrets. 3. If there's no analytics site for the production domain yet, create one, serve the tracker from our own domain as the manual describes, add a goal for each sign-up, purchase or lead, deploy, and run verify. 4. If the site has server or edge code, report bot requests from it: for each request whose user agent matches the prefilter from GET /v1/crawlers, POST /v1/sites/:id/crawls with user_agent, path and status, without waiting for the answer. Keep the key server-side. If the site is static hosting, tell me and skip this. 5. If this project uses an AI visibility tracker (Profound, Peec AI, Otterly, Semrush or another), leave it and its integrations as they are and tell me which one it is. 6. Deploy, then run the setup check (GET https://agentfirststack.com/v1/setup) and fix what it reports. 7. Read GET /v1/sites/:id/ai-search?period=30d. List the pages AI search and answer bots read most, the pages visitors from assistants landed on, and the next_steps. Don't change robots.txt, llms.txt or bot settings without asking me. 8. Write the site id and the env var name (never the key) in AGENTFIRSTSTACK.md, and create a claim link for me with POST https://agentfirststack.com/v1/claim-link. Tell me what you set up, what the report says, what's waiting on me, and give me the claim link.

What AI visibility trackers do

A tracker starts from prompts: the questions your buyers might ask an assistant, such as "best invoicing tool for freelancers". You choose them, or the tool suggests them. It sends them to the assistants your plan covers on a schedule, keeps the answers and scores them: whether your brand is named, where, in what tone, against which competitors, and which pages the answer cites. In the vendors' own words, as of 7 Nov 2026:

Profound runs structured prompts, "analyzing where and how your brand appears in responses", and tracks citations, sentiment, ranking and competitive presence (pricing FAQ).
Peec AI reports "Your visibility, position, sentiment, and share of voice, tracked daily against competitors", plus the sources answers cite (pricing).
Otterly describes its AI Search Analytics as a way to "Track brand mentions and citations across AI engines" (pricing).
Semrush's AI Visibility Toolkit runs common queries "through LLMs like ChatGPT and Google's AI Mode" and analyzes how your brand is mentioned (knowledge base).

That's what people mean by LLM visibility: how present a brand is in the answers language models give. A tracker's numbers are a sample of it: the prompts you picked, run in the assistants, countries and languages your plan includes. Answers can name you without anyone clicking through to your site, and only a tracker sees those.

Trackers have started measuring the site side too. Profound's Agent Analytics says it tracks "when, how often, and which AI bots access your content", from server logs rather than JavaScript, and "Differentiates real AI bots from spoofed crawlers" (Agent Analytics). Peec AI's pricing page lists crawl insights ("Which AI bots hit which pages, how often"), a robots.txt crawlability audit and AI referrals on every plan. Otterly includes Agent Analytics on its Standard and Premium plans. Semrush's toolkit has a site audit that looks for issues that might stop AI bots crawling your content. Their crawler data comes through integrations with your CDN or host (Cloudflare, Vercel, WordPress and others) or uploaded log files.

What they cost

From each vendor's pricing page, as of 7 Nov 2026. "Prompts" are the questions tracked; each runs in each assistant the plan covers.

Tracker	Price	Prompts	Assistants covered	Site-side features
Profound	Trial free; Enterprise "Custom"	Trial: "50 prompts run daily for 7 days"; Enterprise: tailored	Trial: ChatGPT, Gemini, Google AI Overviews; Enterprise: up to 9	Agent Analytics, listed on the Trial with unlimited domains
Peec AI	Starter $95, Pro $245, Advanced $495 a month; Enterprise custom; 15% off annual; free trial	50, 150, 350	Choose 3 of ChatGPT, AI Mode, AI Overviews, Microsoft Copilot, Gemini, Naver AI; Enterprise up to 13, with Perplexity and Claude	Crawl insights (4.5M to 31.5M bot visits by plan), robots.txt audit, AI referrals
Otterly	Lite $29, Standard $189, Premium $489 a month; 15% off annually; Enterprise custom; free trial	15, 100, 400	ChatGPT, Google AI Overviews, Perplexity, MS Copilot; Claude, Google AI Mode and Gemini as paid add-ons	Agent Analytics on Standard (200k events a month) and Premium (1M)
Semrush AI Visibility Toolkit	$99 a month; no free trial	25; 50 more for $60 a month	ChatGPT and Google AI Mode named for prompt tracking	AI search checks in Site Audit for up to 100 pages

Peec AI's page showed prices in dollars when we checked.

What AgentFirstStack measures instead

AgentFirstStack's AI search is part of its web analytics, and it measures what happens on your side:

What AI crawlers read. Your server reports bot requests (POST /v1/sites/:id/crawls, only for user agents matching the prefilter from GET /v1/crawlers), and each is classified: AI answers (ChatGPT-User, Claude-User, Perplexity-User), AI search (OAI-SearchBot, Claude-SearchBot, PerplexityBot), AI training (GPTBot, ClaudeBot, CCBot) and search engines. You see which pages each reads, which got errors, and what they asked for of robots.txt, llms.txt and the sitemap.
The visitors assistants send. Visits referred by ChatGPT, Perplexity, Claude, Gemini, Copilot and others, where they land, and what they do next to every other channel: engagement, conversions, conversion rate and revenue, with your own goals and payments.
What to change. GET /v1/sites/:id/ai-search (MCP get_ai_search) ends with next steps: a robots.txt that keeps out a search or answer bot (checked against the site's own robots.txt), pages that fail for crawlers, a missing llms.txt, pages bots read that send no one.

It's free on the hosted service within fair-use limits, open source (AGPL-3.0, browser scripts MIT) and self-hostable on Cloudflare. Your agent sets it up with one request for a key and reads it over REST or MCP. What it doesn't do: run prompts, count mentions or citations in answers, score sentiment, or compare you with competitors. It classifies crawlers by user agent without checking their IPs, and it can't see an answer that names you if nobody clicks.

Side by side

	AI visibility trackers	AgentFirstStack AI search
Question it answers	What do assistants say about my brand, and whom do they cite?	What do AI crawlers read on my site, and what do the visitors assistants send do?
How it measures	Runs prompts in the assistants on a schedule and scores the answers	Your server reports bot requests; the tracker on your pages records visits
What it covers	The prompts, assistants and countries your plan pays for	Every bot request your server reports and every visit the tracker sees; answers with no click are invisible to it
Competitors	Share of voice, position and citations against competitors	No
AI crawlers	Profound, Peec AI and Otterly (from Standard) through CDN, host or log integrations; Profound says it filters spoofed bots	A few lines of server or edge middleware; classified by user agent, no IP check
Visitors and conversions	Profound and Peec AI report visits from AI assistants	Visits, conversions, revenue and lift against every other channel, with your goals and payments
robots.txt	Peec AI audits it against 40+ bots; Semrush's site audit checks crawler access	Read for AI search and answer bots that stopped at robots.txt, in the next steps
Setup	An account, prompts to track, and an integration for crawler data	`POST /v1/signup`, the tracker, the crawler middleware and goals, done by your agent
Agent access	Peec AI: MCP on every plan; Otterly: API and MCP from Standard; Profound: API on Enterprise	REST and MCP on every account
Price	Paid self-serve plans $29 to $495 a month; enterprise plans on request; Profound's free trial runs 7 days	Free hosted service with fair-use limits
Source	Proprietary	AGPL-3.0, self-hostable; browser scripts MIT

Why the two work together

A tracker tells you where you're missing from answers. AgentFirstStack tells you whether the bots behind those answers can read your pages, and whether the visits that follow are worth having. A typical loop:

The tracker shows prompts where competitors are cited and you aren't, and which of their pages get cited.
The AI search report shows whether OAI-SearchBot, PerplexityBot, Claude-SearchBot and the fetchers that answer in real time read your page on that topic, got an error, or stopped at robots.txt.
You fix access, and rewrite the page so it answers the question in its first paragraph, with specifics worth quoting.
The report shows whether the bots read it again and whether visitors from ChatGPT or Perplexity arrive and convert; the tracker shows whether the answers changed.

When to choose which

Use an AI visibility tracker if

You need to know what assistants say about your brand, product or category, and how you compare with competitors there.
Your team reports share of voice, sentiment or citation share in AI answers.
You want to know which questions to write for: Semrush has prompt research, and Profound lists prompt volume searches.
Few visitors arrive from assistants yet, and you want to know whether you're named in answers anyway.
You want crawler data checked against bots' real IPs: Profound says its Agent Analytics does that.

Use AgentFirstStack AI search if

You want to know whether AI crawlers can and do read your pages, and fix what stops them.
You want visitors from assistants measured against your other channels on conversions and revenue, in the analytics you already use for everything else.
You want it free and open source, set up and read by your agent over an API.

Use both if

You're putting real work into AI search: the tracker measures the answers, AgentFirstStack measures the site and the visits.

Running both

There's nothing to move. AgentFirstStack doesn't import tracker data, and it doesn't need to: the tracker keeps running its prompts, and AgentFirstStack starts counting crawls and visits from the day your agent sets it up. If a tracker already gets your crawler logs through a CDN or host integration, you can keep it; AgentFirstStack's crawler reporting is separate code in your own server or middleware.

FAQ

What is LLM visibility?

How often, where and how a brand shows up in the answers that assistants built on large language models give: ChatGPT, Perplexity, Gemini, Copilot, Claude and Google's AI Overviews and AI Mode. AI visibility trackers measure it by running prompts and scoring the answers for mentions, position, sentiment and the sources cited.

Does AgentFirstStack track brand mentions in ChatGPT or Perplexity?

No. It doesn't run prompts through any assistant, so it can't say how often answers mention or cite you, or how you compare with competitors there. It measures your own site: which AI crawlers read which pages, which assistants send visitors and how those visitors convert, and what to change.

How much does an AI visibility tracker cost?

As of 7 November 2026: Otterly from $29 a month for 15 prompts, Semrush's AI Visibility Toolkit $99 a month for 25 prompts, Peec AI from $95 a month for 50 prompts, and Profound a free 7-day trial of 50 prompts with Enterprise priced on request. Higher plans run to $489 (Otterly) and $495 (Peec AI) a month.

Do AI visibility trackers also show AI crawlers?

Several do. Profound's Agent Analytics reads server or CDN logs, Peec AI lists crawl insights, a robots.txt audit and AI referrals on its pricing page, and Otterly includes Agent Analytics from its Standard plan. Semrush's toolkit has a site audit for AI crawler access rather than crawler logs.

Can I see visitors from ChatGPT in Google Analytics?

Yes. GA4's default channel group now has an AI Assistant channel for sources such as ChatGPT, Gemini, DeepSeek, Copilot and Grok (Google's help page, as of 7 Nov 2026). GA4 doesn't show crawlers: it excludes known bots and doesn't report how many it excluded (known bot-traffic exclusion).

Can AgentFirstStack import data from my AI visibility tracker?

No. There's no import for tracker data, and there's nothing to move: the tracker keeps measuring answers, and AgentFirstStack starts measuring your site and its visitors from the day it's set up.

Product names belong to their owners. AgentFirstStack isn't affiliated with Profound, Peec AI, Otterly, Semrush or Google.