Research · market brief · Updated
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. 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 Found by AI report
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. 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 Vendors change pricing; if a cell is out of date, tell us and we will correct it.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.
| Tool | Entry price | Engines measured at that price | What sits behind a higher tier | Source |
|---|---|---|---|---|
| Profound | Free trial only; paid tiers by sales call | 10 prompts run once, ChatGPT only | Enterprise by sales call, up to nine Answer Engines, tracked daily. The pricing page no longer shows a self-serve paid tier: the US$99 Starter and US$399 Growth tiers it carried on 9 September 2026 are off that page, though the Answer Engine Insights FAQ still describes Starter at 50 monthly prompts and Growth at 100. The API is listed as None on the trial and available on Enterprise; the hosted MCP server's plan gating is still not stated. | Vendor source ↗ |
| Peec AI | US$80/mo billed annually (US$95 monthly) | Any 3 of ChatGPT, AI Mode, AI Overviews, Copilot, Naver AI and Gemini, 50 prompts, one country, refreshed daily; an extra model is US$30/mo. Perplexity is no longer on this tier, it sits on Enterprise and the add-on list | 150 prompts at US$205/mo billed annually, 350 at US$420/mo billed annually, still 3 models each and 3 countries; Claude, Grok and DeepSeek sit on the Enterprise tier with up to 13 models, by sales call; the REST API is Enterprise only, the MCP server works on any account | Vendor source ↗ |
| Otterly.AI | US$29/mo | ChatGPT, Google AI Overviews, Perplexity and Copilot, 15 prompts, refreshed daily, with brand sentiment; US$25/mo billed yearly | Gemini and AI Mode are US$9 a month each as add-ons on the entry plan, Claude US$29; API, MCP and agent analytics start at US$189/mo (US$160 billed yearly) | Vendor source ↗ |
| Surfer | US$82/mo, billed yearly, “AI Search Analytics” | 50 prompts refreshed daily across ChatGPT, Gemini, Google AI Overviews, AI Mode and Perplexity; Mention Gap, Brand Sentiments and MCP included; no Claude, Grok or DeepSeek | The content tools start at US$99/mo billed yearly with 25 prompts refreshed weekly on ChatGPT only, without sentiment or MCP; daily tracking with the content tools is US$182/mo billed yearly; the API is on Peace of Mind only. | Vendor source ↗ |
| Ahrefs Brand Radar AI | From US$50/mo standalone for Custom Prompts; US$199/mo per platform index | Basic includes 2,500 custom prompt checks a month. The selected platforms and checking frequency determine allowance use; Claude consumes eight checks per update. Every paid Ahrefs plan except Starter includes a monthly custom-prompt allowance, 150 checks a month on Lite. | Custom Prompts Growth is US$100/mo for 7,000 checks and Scale US$250/mo for 25,000. All platform indexes cost US$699/mo, including 2,500 custom checks a month. Custom prompts run daily, weekly or monthly; the index refreshes monthly on a 90-day window. API and MCP need a Lite plan or higher. | Vendor source ↗ Additional source ↗ Additional source ↗ |
| Semrush | US$99/mo per domain, billed annually (AI Visibility toolkit on its own); US$165.17/mo Starter inside Semrush One | 25 custom prompts on the standalone AI Visibility toolkit. Inside Semrush One, 50 prompts a day across ChatGPT Search, Google AI Mode, AI Overviews and Gemini, with Perplexity in the weekly Brand Performance report | 100 prompts at US$248.17/mo billed annually, 200 at US$455.67/mo billed annually, inside Semrush One. The AI Visibility toolkit now has its own pricing page rather than being sold only inside the bundle. | Vendor source ↗ Additional source ↗ Additional source ↗ |
| OpenSEO | US$0 self-hosted plus your own DataForSEO usage, or US$10/mo hosted with US$10 of usage included | The brand lookup covers ChatGPT and Google AI Overviews; the Prompt Explorer runs a prompt you type across ChatGPT, Claude, Gemini and Perplexity and shows each answer with its citations | Nothing paywalled; ChatGPT brand checks cost about US$1.09 each in usage, and you do the setup. Its MCP covers keyword, SERP, domain, backlink and rank-tracking tools, not the AI-visibility data. Our comparison: /compare/openseo | Vendor source ↗ Additional source ↗ Additional source ↗ Additional source ↗ |
| Lettertrace | US$0 plus your own API keys | Claude, ChatGPT, Gemini, Google AI Overviews and Perplexity Sonar on your keys | Nothing paywalled; the cost is the provider tokens you pay for and the setup you do. MIT licensed, with an MCP server, a REST API, sentiment per answer and optional repeat asks per prompt. | Vendor source ↗ |
| Are you found by AI? | US$79/mo, Stay Found | All 7 engines: ChatGPT, Claude, Gemini, Perplexity, Grok, DeepSeek and Google AI Overviews, up to 25 buyer questions re-measured every week | Nothing. One plan, the whole board, every week. | Vendor source ↗ |
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. 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.
| Capability | Found by AISource 1 ↗ · Source 2 ↗ · Source 3 ↗ | ProfoundSource 1 ↗ · Source 2 ↗ · Source 3 ↗ | PeecSource 1 ↗ · Source 2 ↗ · Source 3 ↗ | OtterlySource 1 ↗ · Source 2 ↗ · Source 3 ↗ | SurferSource 1 ↗ · Source 2 ↗ · Source 3 ↗ · Source 4 ↗ | Ahrefs Brand RadarSource 1 ↗ · Source 2 ↗ · Source 3 ↗ | SemrushSource 1 ↗ · Source 2 ↗ · Source 3 ↗ | OpenSEOSource 1 ↗ · Source 2 ↗ · Source 3 ↗ · Source 4 ↗ | LettertraceSource 1 ↗ |
|---|---|---|---|---|---|---|---|---|---|
| Engines at the entry price | All 7: ChatGPT, Claude, Gemini, Perplexity, Grok, DeepSeek, Google AI Overviews | ChatGPT only (US$99); 3 at US$399; up to 9 by sales call | Any 3 of 6 (ChatGPT, AI Mode, AI Overviews, Copilot, Perplexity, Gemini), 50 prompts; an extra model 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 | Brand lookup: ChatGPT and Google AI Overviews; Prompt Explorer: ChatGPT, Claude, Gemini, Perplexity (release notes) | 5 on your own keys, incl. Claude |
| Claude, Grok, DeepSeek covered | Yes, all three | Claude, Grok, DeepSeek listed; not on the US$99 tier | 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; Grok and DeepSeek not in the README | Claude yes; Grok and DeepSeek not listed |
| Refresh | Weekly re-measure of up to 25 buyer questions; daily scan of the web for new mentions | 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 | On-demand lookups metered per check; scheduled refresh not in the README | 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 | Every answer kept word for word, 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; word-for-word storage not stated | Prompt Explorer shows each model's answer for a prompt you run, with citations when returned; stored history not in the README | 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: "what your brand sentiment is"; on the US$82 plan and above, not on Standard | Not stated | 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 Starter and Growth; world heat map; personas | By country: 1 on Starter, 3 on Pro and Advanced, unlimited on Enterprise; Gemini from the United States only | 65+ countries incl. Australia, country level | 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 | Not in the README (country pickers are for keyword research and rank tracking) | 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 | 459M+ prompt index (the pricing page says 475M+) | Prompt research with volume, difficulty and intent | Keyword research | 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: how ChatGPT, Gemini, Claude, Perplexity crawl you | 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 Standard; 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; plan gating not stated; REST API Enterprise only | MCP server on any account with a project; REST API Enterprise only; 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 for keyword, SERP, domain, backlink, rank-tracking and Search Console tools; AI-visibility data is not among the MCP tools listed in the docs | 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 | US$79/mo for the seven listed engines and weekly measurement. A card is required for the separate 14-day trial. | US$399/mo billed yearly for 3 engines; all engines by sales call | 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 DataForSEO usage; hosted US$10/mo with US$10 usage included, ChatGPT brand checks about US$1.09 each | 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.
Stay Found includes ChatGPT, Claude, Gemini, Perplexity, Grok, DeepSeek and Google AI Overviews for US$79 a month. Compare that exact set with each tier: some competitors offer other surfaces, including AI Mode and Copilot, that we do not list in our seven.
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.
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.
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.
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.
Profound, Peec, Otterly and Surfer’s tracker refresh daily, as does Semrush prompt tracking. Stay Found normally re-measures weekly and offers a 14-day launch watch. Daily tools can give more observations across a week.
Profound, Peec, Otterly, Surfer, 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.
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.
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 uses ChatGPT and Gemini; the seven-engine measurement belongs to the separate paid service and card-required trial.
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.
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.
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; brand lookup covers ChatGPT and Google AI Overviews, while Prompt Explorer covers ChatGPT, Claude, Gemini and Perplexity. Found by AI's readiness scoring and audit suite are open at github.com/techhorizonlabs/thl-open; the hosted measurement across the seven engines is a service we run.
The 9 September 2026 pricing snapshot, with Ahrefs custom-prompt pricing rechecked on 15 September, runs from US$29 a month (Otterly, four engines) through US$79 to US$99 (Found by AI at US$79 for all seven engines, Peec at US$80 billed annually for any three of six models, Surfer's US$82 tracker billed yearly, Profound at US$99 billed yearly for ChatGPT only), 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. Most tools gate engine coverage by price.
As of 9 September 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's US$99 tier billed yearly tracks ChatGPT only; Ahrefs counts a Claude check as eight checks; Lettertrace covers Claude on your own key, and OpenSEO includes it in Prompt Explorer. Found by AI measures Claude on every paid plan, alongside ChatGPT, Gemini, Perplexity, Grok, DeepSeek and Google AI Overviews.
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.
For agencies, one link runs your whole client list under your logo, every client re-measured weekly, 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 seven-engine service and the evidence it gives you before choosing. Run the free scan and read what AI is telling your buyers today.