SEO & GEO professional · Hands-on review · Updated 2026-11-05
How this review was prepared
Account: an LLM Pulse Starter trial on my own account, started November 5, 2026. Test project: my second site, AcademyGems, a review site for AI courses, with 32 prompts on ChatGPT, Perplexity, Gemini, Google AI Mode and Google AI Overviews, and 5 competitors. What I tested: time to first data, every answer and source it stored, share of voice, citations by domain and page, query fan-outs, AI visibility recommendations, three GEO Writer tasks, the 9-report technical GEO bundle, and the MCP server from Claude Code. I checked its advice against my site and its numbers against the answers themselves. After its second weekly run on November 9, 2026, I fetched the same 32 prompts directly from ChatGPT, AI Mode and AI Overviews twice that morning and compared mentions, citations and the top brands in each answer. Not tested: Growth and Scale features such as sentiment, AI traffic and white label.
Affiliate links on this page may earn me a commission, at no extra cost to you. They don't change what I recommend.
The Verdict
Is LLM Pulse right for you?
LLM Pulse is best for weekly AI visibility reports for small agencies and marketing teams. It tracks 50 prompts on five AI engines for $58 a month and stores every answer with its sources.
Quick answer: which tool fits you
- If you report AI visibility to clients or a team every week
LLM PulseThis reviewStarter $58 a month; Growth $118 for a second brand - If you need numbers that change every day on 50 prompts
Peec AI$95 a month, 50 prompts checked daily - If you want the most daily answers for the money
Promptwatch$95 a month, up to 6,000 answers a month - If you want the cheapest way to start
OtterlyLite $29 a month, 15 prompts checked daily
I tracked 32 prompts for my second site on five AI engines. The first answers arrived 9 seconds after I created the project, every answer is stored with its sources, and its advice found a real gap: AI engines cite my pages without naming the site. But Google AI Overviews described my site as a different company, with the wrong founder and prices, and LLM Pulse counted that answer as a win. On ChatGPT its results matched my own checks on all 32 prompts for brand mentions; on Google's AI answers it matched less often than two of my own checks matched each other.
If you need numbers that move every day, look at Otterly, which I tested for seven days, Peec AI or Promptwatch, which I tracked 24 prompts with for a week, instead: Starter checks each prompt once a week. If you mostly want cheap prompt checks you run from Claude, OpenSEO charged me $0.002 per scheduled answer on its $10 plan, with simpler reports. If you need sentiment, that starts on Growth at $118 a month. It is beginner-friendly through the dashboard and easy to run from Claude through its MCP server. I would pay for it myself.
My rating
4.0 / 5
- Data accuracy (my tests)3.5
Counts mentions correctly and matched my ChatGPT checks on all 32 prompts, but matched Google's AI answers less often than two direct checks matched each other, and scored a wrong answer about my site as a win.
- Features and depth4.0
Full answers with sources, citations by page, query fan-outs, recommendations, GEO Writer and technical audits.
- Value for a small business4.0
50 prompts on five engines for $58 a month, checked weekly; daily checks cost more.
Better here: Otterly (daily from $29)
- Ease of use3.5
Fast setup, but the overview opens on brand prompts only and share of voice hides its sample size.
- AI search features4.5
MCP server with read and write access on every plan, including the trial.
- Trial and billing3.0
14 days, but a card is required, the plan starts automatically and there are no refunds by default.
I rate only tools I have used myself, on criteria from 1 to 5, and leave out any I couldn't measure. The overall score is my own judgement, not an average. Rated 2026-11-09.
Compare it: LLM Pulse alternatives · Best AI visibility tools
Who It's For
Who LLM Pulse is for (and who it isn't)
✓ Best for
- +Weekly AI visibility reports for small agencies and marketing teams. Unlimited team members, full answers with sources, and an MCP server on every plan. Scale adds the projects an agency needs.
- +Small businesses and consultants new to AI visibility. The dashboard is beginner-friendly and its recommendations say what to do next; you need interest in the topic, not technical skills.
- +Marketing teams that review AI visibility weekly. Starter covers 50 prompts on five engines, checked every Monday, for $58 a month.
- +Anyone who works through Claude or ChatGPT. The MCP server reads and writes projects, prompts, answers and reports, including on the trial.
× Consider something else if
- −You want the cheapest way to start. Otterly's Lite plan costs $29 a month and checks daily, with 15 prompts.Try Otterly →
- −You need daily tracking on 50 prompts. Peec AI's Starter checks 50 prompts daily for $95 a month. LLM Pulse's daily Starter costs $94.Try Peec AI →
- −You want the largest daily sample for the money. Promptwatch's Essential plan checks 50 prompts daily for $95 a month, up to 6,000 answers a month.Try Promptwatch →
What You Can Do
What can LLM Pulse do?
What it's like to use
The interface is clear and well organised. I found the MCP connection easily, connected it to Claude Code and did most of my work in LLM Pulse through Claude, so it suits AI-native work well. It works just as well on its own through the dashboard: it gives clear recommendations on what to do next to improve your AI visibility, and it is beginner-friendly. You do need some interest in the topic to get value from it, because the numbers only help if you read the answers behind them and act on them. In my view it fits small businesses, consultants and small agencies. What may surprise you at first is how much there is to fix on a site. Take it in small pieces, or hand everything that can be fixed to your AI assistant straight away, and check the suggestions before you publish. I would pay for it myself.
Setup and how fast the first data arrives
I created a project for AcademyGems with 12 prompts and 5 competitors at 11:50 on November 5, 2026, and later added 20 more prompts. The first answer arrived 9 seconds later, and all 60 answers for the first 12 prompts were in within about 3.5 minutes. After that the plan runs on a weekly schedule: every prompt is asked once per engine, and the next check is on Monday. There is no button to rerun a prompt sooner. After the first run the dashboard showed 32 prompts, 160 answers, a mention rate of 11.9% and a citation rate of 25%, and it says plainly that trend charts appear after the second run. Two things to watch. The overview opens filtered to prompts that name your own brand, where my mention rate read 95%; across all prompts it was 11.9%. And share of voice showed my site at 65.5% against Class Central, built on 29 mentions, with no sample size on screen. The API does report a confidence level for the same number ("low"), so the data is there; the dashboard just doesn't show it.
Time to first answer
- Result
- 9 seconds
All answers for the first 12 prompts
- Result
- About 3.5 minutes
Answers per prompt per engine
- Result
- 1 per weekly run
Mention rate, all 32 prompts
- Result
- 11.9%
Mention rate, prompts that name the brand
- Result
- 95%
Citation rate
- Result
- 25%
| First run, November 5, 2026 | Result |
|---|---|
| Time to first answer | 9 seconds |
| All answers for the first 12 prompts | About 3.5 minutes |
| Answers per prompt per engine | 1 per weekly run |
| Mention rate, all 32 prompts | 11.9% |
| Mention rate, prompts that name the brand | 95% |
| Citation rate | 25% |
AcademyGems, 32 prompts on 5 engines, US, English.


Does it catch an AI answer that gets your brand wrong?
This is the finding that matters most for a small business. For "Is AcademyGems worth it for finding AI courses?", Google AI Overviews said AcademyGems was founded by Andrei Neagoie, bundles about 170 courses and costs about $25 a month or $1,299 for life. That is Zero To Mastery, one of the platforms AcademyGems reviews, not AcademyGems. ChatGPT, Perplexity, Gemini and AI Mode described the site correctly. LLM Pulse stored the wrong answer in full, with a source chip on every claim, which is how I found it. But it scored the answer as a win: brand mentioned, position 1, cited, sentiment positive in the list. Nothing flagged it, and no alert fired. Meanwhile Gemini's answer, which was correct but cautious, was the one marked negative. GEO Writer, asked how to improve the site's AI visibility, flagged Gemini's caution and missed the Overview's error. On the next weekly run, November 9, the same AI Overview called AcademyGems a review platform but still named Andrei Neagoie as its founder and listed Zero To Mastery's prices, and LLM Pulse again recorded a mention at position 1 with my own page cited first. My own direct checks that morning got the same wrong answer twice. Read the answers to your brand prompts yourself; a positive score doesn't mean the answer is right.


Do its numbers match the real AI answers?
On November 9, 2026, LLM Pulse ran its second weekly check between 00:35 and 02:05 CET. The same morning I fetched the same 32 prompts directly from ChatGPT, Google AI Mode and Google AI Overviews, twice, a few minutes apart. The two direct runs show how much the answers move on their own, so LLM Pulse can only be judged against that. Its counting is right: whether it marked my site as mentioned matched the answer text it stored in all 160 answers, and the cited flag missed one source in 160. ChatGPT was the closest match: brand mentions agreed on all 32 prompts, the same as my two direct runs. Google's engines matched less well. For AI Overviews, LLM Pulse found my site cited in 4 answers, while my two later checks found 10 and 8. My checks ran about six hours after LLM Pulse's, so part of the gap is Google changing its answers, and that is the point: on Starter you get one snapshot a week. Treat any single week's number as a range, not a fixed score.
ChatGPT: site mentioned
- LLM Pulse vs my checks
- 100%
- My two checks vs each other
- 100%
ChatGPT: site cited
- LLM Pulse vs my checks
- 81% to 88%
- My two checks vs each other
- 94%
AI Mode: site mentioned
- LLM Pulse vs my checks
- 94%
- My two checks vs each other
- 100%
AI Mode: site cited
- LLM Pulse vs my checks
- 84% to 88%
- My two checks vs each other
- 97%
AI Overviews: site mentioned
- LLM Pulse vs my checks
- 88% to 91%
- My two checks vs each other
- 97%
AI Overviews: site cited
- LLM Pulse vs my checks
- 81% to 88%
- My two checks vs each other
- 94%
Top 3 brands named, average overlap
- LLM Pulse vs my checks
- 46% to 64%
- My two checks vs each other
- 54% to 66%
| Agreement on 32 prompts, November 9, 2026 | LLM Pulse vs my checks | My two checks vs each other |
|---|---|---|
| ChatGPT: site mentioned | 100% | 100% |
| ChatGPT: site cited | 81% to 88% | 94% |
| AI Mode: site mentioned | 94% | 100% |
| AI Mode: site cited | 84% to 88% | 97% |
| AI Overviews: site mentioned | 88% to 91% | 97% |
| AI Overviews: site cited | 81% to 88% | 94% |
| Top 3 brands named, average overlap | 46% to 64% | 54% to 66% |
Citations and sources: what you can act on
The source data is where LLM Pulse earns its price. Citations by domain show which sites each engine relies on: for my prompts, AI Overviews leaned on YouTube and Reddit, while Perplexity cited my own site most. Citations by page list my cited pages with their HTTP status and a saved copy of each page as the engine saw it. Query fan-outs show the searches an engine ran behind a prompt, which tells you what a page has to rank for. The most useful pattern: my site was cited as a source in 25% of answers but named in only 11.9%, and only on prompts that already contained its name. Perplexity used my review pages as its first source for "Is DeepLearning.AI worth it?" and "Zero To Mastery review" without naming the site. That is a fixable problem, and LLM Pulse is what showed it to me.



Recommendations, GEO Writer and audits
I ran everything the Starter plan allows: four recommendation sets, all three of the month's GEO Writer tasks, and the 9-report technical GEO bundle. It is fast, from 10 seconds to 90 seconds per job, and the AI visibility recommendations are tied to my own prompts and cited pages rather than a generic checklist. The best output was a GEO Writer audit of my About page: it traced the AI Overview mistake to a listing card on my homepage whose summary started "A subscription academy founded by..." without naming Zero To Mastery, and drafted a definition and FAQ to fix it. I applied that fix the same day. Four days later, Google had picked up the definition but not dropped the wrong founder and prices, so a fix like this takes more than one weekly run to show. But every output needed checking. One recommendation said no page on my site covers vibe coding; there are several. GEO Writer suggested publishing an llms.txt file that already exists, and called a real Skool member rating a contradiction. Neither GEO Writer task marked the claims I asked it to flag for checking. The robots.txt report listed 19 critical issues, all caused by normal disallows of redirect and admin paths. And two brand-building recommendations proposed paid placements, one asking the publisher to leave a sponsored article without a sponsored label, which breaks Google's link spam rules. Use the advice as a list of leads, and check each one against your site before you act.
AI visibility recommendations (8)
- What I found
- Tied to my prompts; 1 built on a false premise
GEO Writer, About page audit
- What I found
- Found the cause of the wrong AI answer
GEO Writer, open question
- What I found
- Found 'cited, not named'; missed the wrong answer
Brand building recommendations (8)
- What I found
- 6 outreach leads, 2 paid placements
Technical GEO bundle (9 reports)
- What I found
- Score 18 to 100; 19 'critical' robots.txt issues were normal disallows
| Advice feature, November 5, 2026 | What I found |
|---|---|
| AI visibility recommendations (8) | Tied to my prompts; 1 built on a false premise |
| GEO Writer, About page audit | Found the cause of the wrong AI answer |
| GEO Writer, open question | Found 'cited, not named'; missed the wrong answer |
| Brand building recommendations (8) | 6 outreach leads, 2 paid placements |
| Technical GEO bundle (9 reports) | Score 18 to 100; 19 'critical' robots.txt issues were normal disallows |
Starter allows 3 GEO Writer tasks a month and 200 manual audit runs a day.



What the Starter plan covers, and what it doesn't
Starter covers one project, 50 prompts and 10 competitors on five engines: ChatGPT, Perplexity, Gemini, Google AI Mode and Google AI Overviews. Claude, Copilot, Grok, DeepSeek, Meta AI and others are paid add-ons. Each prompt is checked once per engine every week, so the plan's real allowance is 263 answers a week (50 prompts times 5 engines, plus 5%); a new prompt spends one answer per engine on its first run, and deleting it doesn't give the answers back. Team members are unlimited on every plan. Sentiment, social and reputation views are locked to Growth and above, although sentiment labels still appeared in my Starter answer list. Competitors per project rise from 10 on Starter to 15 on Growth and 20 on Scale. For an agency, the limit that matters is projects: Starter has one, Growth two and Scale five, and an extra project costs about $60 a month. White label is on Enterprise only.

Use it through Claude with the MCP server
The official MCP server works on every plan, including the trial, and connects with OAuth. From Claude Code I read the account limits, projects, prompts, answers, share of voice and sources, added 20 prompts, launched recommendations, GEO Writer tasks and the technical GEO bundle, and read the results back. Everything I did in this review through the dashboard I could also do through Claude. The API reports what the dashboard leaves out, such as a confidence level and margin of error on share of voice. Write actions spend the same quotas as the app, so ask your assistant to show you what it will create before it runs. The connection itself was easy: I found the MCP settings quickly, added the server to Claude Code, and ran most of my LLM Pulse work from there.
Works With
Which website platforms does LLM Pulse work with?
Tracks AI answers by prompt; it does not need to connect to your website.
LLM Pulse also has official integrations with:
Shopify
How each connection works (1)›
ShopifyApp store listing · Shopify admin app shows AI visibility per product and can write page edits back. Source
Checked against LLM Pulse's own docs and official app stores on 22 October 2026.
AI, API & MCP
How LLM Pulse fits an AI-agent workflow
AI agent workflow
Run LLM Pulse from Claude with the MCP server
Prompts that worked for me: "Show this week's share of voice per AI engine with the sample size." "List the answers that mention my brand and quote what they say about it." "Which of my pages are cited without my brand being named?" "Which domains does each engine cite most for my prompts?" Ask for the confidence level whenever you read a share of voice, and read the answers to your brand prompts, not just the scores.
Approval boundary: The MCP server can create prompts, competitors, audits and content tasks, and those spend your plan's quotas. Deleting a project can't be undone.
MCP
officialOfficial MCP server on every plan
Remote server with OAuth, available on every plan including the trial. It reads and writes projects, prompts, competitors, answers, citations, recommendations, GEO Writer tasks and GEO audits. Some tools follow plan limits, such as sentiment on Growth and above.
LLM Pulse →API
officialREST API on Scale and above
The REST API and API-key access start on Scale. The MCP server needs no API key and works on every plan.
LLM Pulse pricing →Pricing
How much does LLM Pulse cost?
Starter costs $58 a month (49 EUR) for one project and 50 prompts checked weekly on five AI engines. Growth is $118 a month (99 EUR) and Scale $354 (299 EUR). LLM Pulse bills in euros, so the dollar prices are its own daily conversion, and VAT comes on top.
| Plan | Monthly |
|---|---|
| Starter · 50 prompts, 1 project | $58/mo |
| Growth · 150 prompts, 2 projects | $118/mo |
| Scale · 450 prompts, 5 projects | $354/mo |
| Enterprise | Custom |
Trial: 14 days on the weekly Starter, Growth or Scale plans, card required, and the subscription starts automatically when the trial ends; daily plans start paid. Daily tracking costs more on every tier: daily Starter is $94 a month (79 EUR). Extra prompts cost 1 EUR (about $2) per prompt a month, an extra project about $60 a month, and extra AI engines from about $10 a month each. Paying yearly saves about 17%.
Daily checks cost 50 to 60% more
The same plans with daily checks: Starter $94, Growth $177 and Scale $532 a month (79, 149 and 449 EUR). Daily plans have no free trial.
Weekly is enough for most small teams
AI answers change from run to run, so one week's number is a sample. Read the trend over several weeks.
Count answers, not prompts
50 prompts on five engines is 250 answers a week. Each extra engine multiplies the count.
Growth for sentiment and a second project
Sentiment and AI traffic start on Growth. Scale adds five projects, the REST API and Looker Studio, which an agency with several clients needs.
The trial needs a card
14 days on the weekly plans; the paid subscription starts automatically when the trial ends.
Alternatives
When I'd choose something else
Three different buying priorities, and what I'd pick instead for each one.
Pick it for the cheapest daily checks
Otterly
Daily checks from $29 a month, for 15 prompts. 50 prompts means Standard at $189 a month (100 prompts).
Trade-off: LLM Pulse wins when you want 50 prompts on five engines for $58 and can live with weekly checks.
Explore Otterly for this use case →Pick it for daily checks on 50 prompts
Peec AI
50 prompts checked daily for $95 a month, up to 4,500 answers a month. LLM Pulse's daily Starter costs $94.
Trade-off: LLM Pulse wins on weekly price ($58) and on its MCP server, which works on every plan.
Explore Peec AI for this use case →Pick it for the most daily answers
Promptwatch
50 prompts checked daily for $95 a month, up to 6,000 answers a month, the most of these three at that price.
Trade-off: LLM Pulse wins when weekly reports are enough and $58 a month matters more than daily data.
Explore Promptwatch for this use case →Keep going
- Compare LLM Pulse against the field: 14 Best AI Visibility Tools in 2026: Real Prices Compared, Best AI Visibility Tools for Agencies
- Put it to work: How to Track Your Brand in AI Answers: Free Data First, Then a Tracker
- Put it to work: How to Rank in AI Search: A Practical GEO Guide for 2026
Downsides
Limitations to consider before subscribing
Wrong answers can count as wins
A wrong AI Overview about my site scored as mentioned, position 1 and positive.
Weekly checks on standard plans
One answer per prompt per engine each Monday, and no way to rerun sooner.
Small samples look certain
Share of voice shows no sample size in the dashboard, only in the API.
Advice needs checking
False premises, an existing llms.txt flagged as missing, and paid link suggestions.
Sentiment starts on Growth
Starter has no sentiment view, and the labels it does show disagree with the detail view.
My Recommendation
My recommendation: use the trial to read your brand answers
Start the 14-day trial, add 20 to 30 prompts a customer would ask, including two or three that name your brand, and read every answer to the brand prompts. If the answers and sources show you something to fix, the weekly Starter at $58 a month is a fair price for keeping an eye on it.
Choose Otterly if you want the cheapest daily checks, or Peec AI or Promptwatch if you need daily tracking on 50 prompts; Promptwatch gives the most answers a month for the price.
What to check during the trial
- –Add prompts that name your brand and read each engine's answer in full.
- –Switch the overview to all prompts before you read the mention rate.
- –Check share of voice for the number of mentions behind it.
- –Compare cited pages with named mentions to find pages that are cited but not credited.
- –Check any recommendation against your own site before you act on it.
FAQ
Frequently asked questions about LLM Pulse
Starter costs $58 a month (49 EUR) for one project and 50 prompts checked weekly on five AI engines, or $49 a month billed yearly. Growth is $118 and Scale $354 a month. Daily tracking costs more, from $94 a month. LLM Pulse bills in euros and VAT comes on top.
Yes, 14 days on the weekly Starter, Growth and Scale plans. A card is required, and the subscription starts automatically when the trial ends unless you cancel before then. Daily plans have no trial, and there is no free plan.
Weekly on the standard plans: each prompt is asked once per engine every Monday. New prompts get their first answers right away; mine arrived within seconds. Daily tracking is available on separate, more expensive plans.
Every plan includes ChatGPT, Perplexity, Gemini, Google AI Mode and Google AI Overviews. Claude, Copilot, Grok, DeepSeek, Meta AI and others are paid add-ons.
Its counting is accurate, and on ChatGPT it matched my own same-day checks for brand mentions on all 32 prompts. On Google AI Mode and AI Overviews it matched 81% to 94% of the time, while two of my own checks matched each other 94% to 100% of the time, so read one week's number as a range. It stores the full answer and sources, so you can check any number yourself. In my test it counted a wrong AI Overview answer about my site as a positive mention at position 1, so read your brand answers rather than trusting the score.
Yes. It is best for weekly AI visibility reports for small agencies and marketing teams: team members are unlimited, every answer comes with its sources, and the MCP server works on every plan. Starter has one project, so client work starts on Growth (two projects) or Scale (five).
Yes. The official MCP server connects with OAuth on every plan, including the trial, and can read and create projects, prompts, audits and reports. I ran this whole review's data pulls through it from Claude Code.
Otterly's Lite plan costs $29 a month and checks 15 prompts daily. LLM Pulse's Starter costs $58 a month for 50 prompts checked weekly on five engines, with full answers, sources and query fan-outs. Pick Otterly for daily checks on a few prompts, LLM Pulse for a wider weekly sample.
Updates
What changed in this review
- 2026-11-09 Accuracy test added: LLM Pulse's November 9 run compared with the same 32 prompts fetched directly from ChatGPT, AI Mode and AI Overviews, twice. Data accuracy score added.
- 2026-11-05 Rewritten as a hands-on review on a Starter trial: 32 prompts on five AI engines for my second site. Prices re-checked on llmpulse.ai (USD display up about 1%, euro prices unchanged). Rating 4 out of 5 added.
- 2026-09-15 First version, researched from LLM Pulse's website without a hands-on test.
Joonas Rotko
SEO & GEO professional, founder of That Marketing Buddy
I'm a Finnish SEO and GEO professional with 11+ years in digital marketing, including years as a co-owner of a digital marketing agency. On That Marketing Buddy I test and compare SEO and AI-visibility software for entrepreneurs, marketing professionals and agencies.
Sources & References (2)
Spot something out of date?Vendors and readers, send me the correct fact.
Software sites change almost every week, and I want every claim on this page to be factually correct. If you are a vendor or a reader, send me the correct fact and a link to the official page. I will check it and update the page. The opinions and rankings stay mine.

