AI visibility tools, mapped: who measures what, at what price

Research: market brief, updated 3 November 2026

The useful comparison is the job you need done: which engines and locations are measured, how often, what evidence you can inspect, and what help you get next. Are you found by AI? is one of the products below; we also show where another tool may fit better.

Compare capabilities and trade-offs ↓ · See a full Are you found by AI? report

What each published entry tier actually measures

Published pricing read on 9 September 2026. Ahrefs custom-prompt pricing, Surfer's AI Search Analytics package and Peec's model limits were rechecked on 15 September 2026, and Profound's plans, Ahrefs' prompt index and Peec's country settings on 6 October 2026. OpenSEO tracking, costs and MCP access, and Ahrefs’ prompt-index size were rechecked on 13 October 2026. Selected Ahrefs, Semrush, Otterly, Surfer, Peec and OpenSEO pricing and capability cells were rechecked on 3 November 2026. Peec annual dollar prices remain the earlier snapshot and could not be re-verified on that date. Links in the table let you verify the details at the source. Vendors change pricing; if a cell is out of date, tell us and we will correct it.

Swipe the table to compare all columns on a phone.

What each tool does with the answers

Price tells you what a tier costs. This table compares capabilities from each vendor’s published pages as read on 9 September 2026. Ahrefs custom-prompt pricing and Peec’s model limits were updated on 15 September 2026. The Profound, Semrush, Peec and Ahrefs rows were re-read on 22 September 2026; Peec’s own price figures could not be re-verified that day because the pricing page rendered them client-side, so those two prices are as last confirmed on 15 September 2026. Profound’s trial, paid tiers and MCP access, Ahrefs’ prompt-index size, Peec’s location wording and Otterly’s country counts were updated on 6 October 2026. Our own pricing and schedule were updated on 12 October 2026. OpenSEO tracking, costs and MCP access, and Ahrefs’ prompt-index size were updated on 13 October 2026. Selected Ahrefs, Semrush, Otterly, Surfer, Peec and OpenSEO cells were rechecked on 3 November 2026. Peec annual dollar prices remain the earlier snapshot and could not be re-verified on that date. Plan restrictions are included in the cells.

“Not stated” means the pages we read did not establish it. It does not mean the feature is absent. This is a comparison of documented capabilities, not an independent test of each vendor’s accuracy.

Swipe horizontally to read every column on a phone. Choose a tool above for a closer comparison.

Published capabilities: 9 September; selected updates through 3 November 2026

Capability	Are you found by AI?Source 1 ↗ · Source 2 ↗ · Source 3 ↗	ProfoundSource 1 ↗ · Source 2 ↗ · Source 3 ↗ · Source 4 ↗	PeecSource 1 ↗ · Source 2 ↗ · Source 3 ↗ · Source 4 ↗ · Source 5 ↗	OtterlySource 1 ↗ · Source 2 ↗ · Source 3 ↗ · Source 4 ↗	SurferSource 1 ↗ · Source 2 ↗ · Source 3 ↗ · Source 4 ↗	Ahrefs Brand RadarSource 1 ↗ · Source 2 ↗ · Source 3 ↗ · Source 4 ↗	SemrushSource 1 ↗ · Source 2 ↗ · Source 3 ↗	OpenSEOSource 1 ↗ · Source 2 ↗ · Source 3 ↗ · Source 4 ↗ · Source 5 ↗	LettertraceSource 1 ↗
Engines at the entry price	The major AI engines	3 on the free trial, up to 9 on Enterprise	Any 3 of 6 (ChatGPT, AI Mode, AI Overviews, Copilot, Gemini, Naver AI), 50 prompts; an extra model, Perplexity included, is US$30/mo on Starter; up to 13 models on Enterprise, including Claude, Grok and DeepSeek	4 (ChatGPT, AI Overviews, Perplexity, Copilot), 15 prompts on Lite; Gemini and AI Mode US$9 each and Claude US$29 as add-ons	5: Gemini, AI Overviews, AI Mode, ChatGPT, Perplexity on the US$82 plan; the US$99 Standard plan tracks ChatGPT only	Custom Prompts from US$50/mo for 2,500 checks, with platforms and cadence selected by you; Claude uses 8 checks per update. Platform indexes are US$199 each or US$699 for all 6.	Daily prompt tracking on ChatGPT Search, AI Mode, AI Overviews and Gemini; weekly Brand Performance adds Perplexity	Scheduled tracking covers ChatGPT, Gemini, Google AI Overviews, Claude and Perplexity. ChatGPT and Gemini use consumer-site collection; Claude and Perplexity use model APIs.	5 on your own keys, incl. Claude
Claude, Grok, DeepSeek covered	Yes, all three	Claude and DeepSeek on Enterprise only. Its pricing page no longer lists Grok.	All three listed, Enterprise only	Claude as a paid add-on; no Grok or DeepSeek listed	None of the three	Claude on custom prompts only, at 8 checks of allowance per execution; Grok listed but paused; DeepSeek not listed	Not stated	Claude in Prompt Explorer and tracking, at live API pricing; Grok and DeepSeek are not listed in the AI Visibility guide.	Claude yes; Grok and DeepSeek not listed
Refresh	Pro checks up to 25 buyer questions on the schedule you choose; every run uses the quoted account tokens	Daily, all plans	Daily on every self-serve plan; daily or weekly on Enterprise	Daily, all plans	Daily on the US$82 plan, Pro and Peace of Mind; weekly on Standard	Index chatbots monthly on a 90-day window; custom prompts daily, weekly or monthly, metered in checks	Daily prompt tracking; brand performance weekly	Manual checks or scheduled daily, weekly or monthly collection, with a cost preview before starting.	On demand, or scheduled daily or weekly
Asked more than once	Adaptive sampling: normally two answers per chat engine, a third for disagreement or an empty answer; majority vote. One AI Overviews read. Long-absent weekly questions may use one probe. The initial 12-question report has a ceiling of 228 answers.	Daily collection; published plan counts are consistent with one response per prompt per day, but do not establish the sampling algorithm.	Once per model per day ("one chat result per model", pricing FAQ)	Not stated	Yes: multiple queries per day per model, averaged (AI Tracker FAQ, live read 9 Sep)	One execution per check (1 prompt, 1 LLM, 1 location); no repeat stated	Not stated	Not in the README	Optional: 1 to 10 replicates per prompt per run, default 1
The answers themselves	The full text of each answer, with the sources returned	Yes: the AI-generated answer captured for every tracked prompt each day, from the browser	Yes: full chat text kept; the last 100 chats per prompt viewable; all chats exportable	Yes: per-run response text with citations, in the report and the API	Yes: the Responses panel shows "the exact answers generated by AI for a chosen prompt", with Date Range and AI Models filters (Fanout Queries docs, live read 9 Sep)	Raw responses stored for the index and searchable; per-answer display not stated	Snapshot of the actual AI result per tracked prompt; full-answer storage not stated	Saved answers, citations and competitor appearances, with trends and export. Reading saved results uses no credits.	Yes, stored responses on your keys
Searches the engine ran first	Kept where an engine exposes them, beside the answer	"Query Fanouts" tab	Yes: fan-out queries shown per chat, also via API and MCP	Yes: the web search queries per response in the API, and a Query Fan-Out view in Prompt Research	Yes: Fanout Queries dashboard per prompt, with a Responses panel (docs, live read 9 Sep)	Fanout queries shown per prompt	Not stated	Not in the README	Not stated
Competitors named instead of you	Yes, per question, with how often	Share of voice and competitor rankings	Benchmark against your market; every detected brand counted, not only the tracked ones	You plus top 5 competitors	Share of voice and competitor benchmark: "which models skip your brand in their favor"	Yes, across the index	Up to four rivals, topic and prompt gaps	Share of voice against competitors; brands mentioned alongside yours (release notes)	Share of voice against each competitor you track
Sources the engines cite	Yes, per answer	Citations tab with source authority	Yes: every URL referenced, with citations in the text split from sources read	Citation tracking	Cited sources are shown per model in the AI Tracker interface.	Yes	Yes: cited domains and pages per tracked prompt	"AI mentions, citations, and prompts"	Yes
Sentiment or portrayal	Not offered	Sentiment tab	Sentiment	Brand sentiment: every mention scored positive, neutral or negative, with the sentence that triggered it	Yes: the sentiment of the cited sources, on AI Search Analytics (US$82/mo billed yearly), Pro and Peace of Mind; not on Discovery or Standard	Brand sentiment across AI search; the product page does not state plan-level availability	Sentiment and narratives	Not in the README	Yes: sentiment, and whether the answer recommended you
Regions and countries	Country-anchored buyer questions and local suburb comparisons where applicable. Australian registry checks are part of the local workflow.	Prompts by country, city or language (150+ regions, 30+ languages); one region and one language on the trial; world heat map; personas	By country: 1 on Starter, 3 on Pro and Advanced, unlimited on Enterprise	Country-level tracking, including Australia. The pricing page says unlimited countries and languages on every plan; the help page says 70+ countries. Check coverage for your engine and country.	Not stated on the AI Tracker or pricing pages	Custom prompts tracked per location; index prompts asked in the keyword's own country	Country, state, region and city: 68,000+ locations for Brand Performance; 220+ countries for prompt tracking	Choose a market for tracking. Prompt Research currently covers US English; this is separate from tracking coverage.	Not stated
Identity check against a business registry	Australian ABN lookup; verified and suggested matches are labelled separately. Availability depends on finding a matching registry record.	Not stated	Not stated	Not stated	Not stated	Not stated	Not stated	Not in the README	Not in the README
Prompt demand data	Connected Search Console query demand, Google keyword volumes and third-party AI-demand estimates where available. These are different datasets, not our own conversation panel.	"Prompt Volumes" from millions of conversations	AI-suggested prompts with a 1 to 5 relative volume score	Prompt Research tool on all plans, with an Estimated Intent Score in place of volumes	Not stated	500 million+ organic prompts (Ahrefs pricing page, checked 3 November 2026; rounded down from 525 million+)	Prompt research with volume, difficulty and intent	Prompt Research uses DataForSEO questions, mostly from Google People also ask; it does not estimate AI question volume.	Not stated; prompts are generated variations per topic
Crawl or agent traffic analytics	A direct agent read of your site plus our visitor beacon. Connected GA4, PostHog and Plausible can add attributed traffic; this is not a server-log crawler dashboard.	Agent Analytics: which AI systems crawl your site, and whether that becomes real traffic	Yes: Crawl Insights from server logs on all plans (Starter 4M bot visits a month), plus a crawlability audit	Agent Analytics from Standard (US$189/mo), 200k events a month; not on Lite	Not stated	Bot Analytics, free while in beta: AI crawler visits in real time	Site Audit flags which AI crawlers you block; AI Traffic dashboard for ChatGPT and Gemini referrals	Site audit crawler of your own pages; AI-agent traffic analytics not in the README	Not in the README
The fix	Entity-specific schema and llms.txt drafts, plus prioritised readiness changes for review and implementation.	"Agents" and weekly highest-impact work	Recommended actions and agent actions; no written fixes stated	Smart recommendations and content briefs	AI Tracker recommendations are documented. Content Editor starts on Discovery; Audit on Pro. Check which tools your chosen tracker or content plan includes.	Content gaps from fanout queries and citation targets inside the SEO suite; no written fix stated	Recommendations inside the SEO suite	SEO site audit plus agent skills	None, tracking only
Your own AI reads the measurement	MCP reads measured answers, rivals, fixes and traffic. get_ai_traffic already includes connected GA4 and PostHog/Plausible summaries and breakdowns. Dedicated GSC query reads are planned.	Hosted MCP server and REST API on Enterprise	MCP server on any account with a project; REST API on Pro, Advanced, Enterprise and agency plans, not Starter; Looker Studio from Advanced; CSV export	API and MCP from US$189/mo (US$160 billed yearly), 2,000 requests a month each; not on Lite	MCP supports AI Tracker reads on AI Search Analytics, Pro, Peace of Mind and Enterprise. Its current docs exclude prompt management, sentiment, fanout queries and mention recommendations. API on Peace of Mind.	API and MCP on Lite plans and above; standalone Brand Radar not stated	MCP and API in Semrush One plans; the MCP tool list has no AI-visibility tools	MCP researches prompts, configures tracking, previews costs, starts or pauses schedules, and reads or exports saved answers, trends, citations and competitor appearances.	MCP server, REST API and CLI
Method you can inspect	Readiness scoring and audit suite open at thl-open; the hosted measurement is a service we run	Source not published	Metric formulas documented; source not published	OpenAPI spec published; source not	"Proprietary algorithm", its own words	A methodology post; source not published	Source not published	Whole platform MIT	Whole tool MIT
Whole board, one price	Free plan for up to 3 sites. Pro is US$79/month or US$869/year per account, plus GST for Australian customers; 14,000 tokens are added monthly and shared across its sites.	Free 7-day trial on three engines; Enterprise priced by quote for a tailored tracking plan	US$80/mo billed yearly (US$95 monthly) for 3 of 6; Pro US$205, Advanced US$420	US$29/mo (US$25 billed yearly) for 4 engines and 15 prompts; add-ons for the rest	US$82/mo billed yearly (US$95 monthly) for 5 engines, 50 prompts	Custom Prompts standalone: US$50 for 2,500 checks/mo, US$100 for 7,000 or US$250 for 25,000. Platform indexes: US$199 each or US$699/mo for all, with 2,500 custom checks included.	US$165.17/mo billed yearly (US$199 monthly), Semrush One Starter, 50 prompts; AI Visibility Toolkit alone US$99/mo, 25 prompts	US$0 self-hosted plus provider usage, or US$10/mo hosted with US$10 of usage included. The pricing calculator estimates about US$0.002 per prompt per platform. Immediate checks cost more than scheduled checks; Claude and Perplexity use live API pricing. Preview the cost before starting.	US$0 plus your provider usage. The README documents an operator-configurable hosted trial, with a default allowance of 15 engine runs.

Refresh frequency, repeated samples and answer surfaces are different choices. A browser-captured answer and an API answer can differ. More engines or a lower price alone do not establish better accuracy.

Where Are you found by AI? is a strong fit

The major AI engines in the monthly price

Monitoring covers the major AI engines for US$79 a month. Each competitor's tier covers a different set, and some include surfaces we do not measure, such as AI Mode and Copilot.

A method you can inspect

Our core paid question board uses adaptive sampling and a published voting method. Surfer also repeats queries and averages them; repetition is not exclusive to us. Our readiness scoring and audit suite are open source, while the hosted visibility engine is a service we run.

From a website to a usable fix

Start with your website. We infer buyer questions for you to review, keep the engine answers and prepare entity-specific schema, llms.txt and prioritised changes. You do not need to bring provider API keys. Other products also suggest prompts and actions; inspect their output as well as ours.

Local identity and context

Australian business-registry checks and location-specific questions help distinguish the right business and market. A suggested registry match is not a verified one. Profound and Semrush also offer city-level targeting; geographic detail is not exclusive to us.

Public comparisons you can inspect

Our public benchmarks show dated, category-scoped measurements beyond your own account. Use the cohort and method shown on each page; a benchmark is context, not proof that any fix caused a change.

Where another tool may fit better

More frequent default tracking

Profound, Peec, Otterly and Surfer’s tracker refresh daily, as does Semrush prompt tracking. Pro checks run when you choose, with their token cost shown before confirmation. Daily tools can give more observations across a week.

Sentiment and longer answer exploration

Profound, Peec, Otterly, Surfer, Ahrefs, Semrush and Lettertrace describe sentiment features. We do not currently offer sentiment classification. Peec and Surfer also expose richer answer-history navigation than our current console.

Larger demand and traffic products

Profound, Ahrefs and Semrush offer large prompt-demand datasets. Profound and Peec include server-log agent analytics; Otterly offers it from Standard. Our query-demand and attributed-visitor data answer different questions.

The broader SEO workflow

Surfer, Ahrefs, Semrush and OpenSEO offer content, keyword or search-research workflows beyond our visibility measurement and fix plan. For a content team already using those tools, keeping that workflow may matter more than engine count.

Read a full example report or run your free scan before choosing. The free scan checks two questions on two engines. A free account keeps up to 3 sites with monthly checks; Pro checks 25 questions across the major AI engines when you choose.

The free, open-source end of the market

Lettertrace and OpenSEO publish their source under the MIT licence. Self-hosting gives you control, with setup and provider usage costs to account for. Lettertrace also documents a hosted trial and URL-based onboarding. OpenSEO offers a US$10 monthly hosted subscription with US$10 of usage included. These are practical alternatives for someone who wants to own the stack; free software does not make its data or operation costless.

What buyers should check, whichever tool they pick

Inspect an actual answer, check your tier’s engine coverage and cadence, and ask how sampling, citations and partial failures affect the result. Check the path from a finding to a usable change. We keep the answers and publish our readiness method and audit suite; the hosted visibility engine is not open source. Those are separate claims, and both matter when choosing a tool.

Questions about the comparison

Are there free open-source AI visibility tools?

Yes. Lettertrace (MIT-licensed) tracks brand mentions across Claude, ChatGPT, Gemini, Google AI Overviews and Perplexity for free: you supply your own provider API keys, curate the prompts, pay the provider's token costs and self-host or use their instance. It measures visibility and does not write fixes. OpenSEO (MIT-licensed) is an open-source SEO platform with an AI visibility feature: its Prompt Tracking runs your questions on ChatGPT, Gemini, Google AI Overviews, Claude and Perplexity, daily, weekly or monthly. The readiness scoring and audit suite are open at github.com/techhorizonlabs/thl-open; the hosted measurement across the major AI engines is a service we run.

How much do AI visibility tools cost in 2026?

The 9 September 2026 pricing snapshot, with Ahrefs custom-prompt pricing rechecked on 15 September and Profound on 6 October, runs from US$29 a month (Otterly, four engines) through US$79 to US$82 (Are you found by AI? at US$79 for the major AI engines, Peec at US$80 billed annually for any three of six models, Surfer's US$82 tracker billed yearly), to Semrush One from US$165.17 a month billed annually, with a separate US$99 AI Visibility Toolkit listed in its knowledge base and Ahrefs Brand Radar custom prompts from US$50 a month standalone, with AI platform indexes at US$199 each or US$699 for all of them. Profound now offers a free 7-day trial. Agency client workspaces are US$399 a month, and Enterprise is priced by sales call. Most tools gate engine coverage by price.

Which AI visibility tools cover Claude?

As of 6 October 2026, Peec's self-serve tiers list six models and Claude is not among them, with "all models" reserved for its Enterprise tier; Otterly sells Claude as a paid add-on from US$29 a month; Profound tracks Claude on Enterprise only; Ahrefs counts a Claude check as eight checks; Lettertrace covers Claude on your own key, and OpenSEO includes it in Prompt Tracking (rechecked 3 November 2026). Are you found by AI? measures Claude on every paid plan, alongside the other major AI engines.

What should I check before buying an AI visibility tool?

Check the answers behind the score, the engines and locations included at your tier, the sampling and refresh method, and how you can inspect past results or request a new measurement. Also compare citations, sentiment, integrations and the work needed to turn a finding into a fix.

If you run an agency

For agencies, one link runs your whole client list under your logo, client checks run on the schedule you choose from shared account tokens, and each person on your team gets a scoped key so their AI reads only the clients they run; ask for an agency rate at /agencies.

Compare our service and the evidence it gives you before choosing. Run the free scan and read what AI is telling your buyers today.