AI visibility monitoring tools: a comparison of seven platforms
Seven platforms now measure whether a brand is cited by ChatGPT, Gemini, Perplexity and Copilot. They cost between US$179 and US$489 per month and differ on almost everything that matters. Only two publicly document coverage for Brazil and Portuguese: Otterly.AI and Semrush. The other five publish nothing verifiable on the subject, which makes language the tiebreaker before price for any non-English operation. Every figure below was read at the source on September 15, 2026.
How was this comparison made?
Every figure was read on the vendor's own pricing page or product documentation, on September 15, 2026. Nothing here comes from a press release, a third-party roundup or memory.
Where the public source does not answer, the cell says not verifiable. That is the finding, not a hedge. A vendor who does not publish language coverage is asking you to carry the risk of finding out later.
Three method caveats, stated up front:
- List prices change without notice. This comparison carries a date and is only valid for that date.
- Annual and month-to-month pricing are not the same number. Where a vendor publishes only the annual rate, the table says so.
- A prompt, a query and a check are different units. Comparing raw counts across vendors is misleading.
What does each tool cost today?
The table uses each vendor's entry plan that is actually usable professionally. Plans between US$29 and US$99 exist and cover a single brand.
| Tool | Plan | Month to month | Annual (per month) | Volume included |
|---|---|---|---|---|
| Otterly.AI | Standard | US$189 | US$160 | 100 prompts |
| Scrunch | Starter | US$300 | US$250 | 350 prompts + 1,000 industry |
| Profound | Growth | not published | US$399 | 100 prompts |
| Peec AI | Pro | not public | not public | 150 prompts |
| Semrush | Pro+ | US$299 | US$248 | 100 prompts per day |
| Ahrefs Brand Radar | Lite + Basic pack | US$179 | — | 2,500 checks |
| SE Ranking | Core + AI Search add-on | US$218 | US$174 | 450 prompts |
Sources: each vendor's official pricing page, retrieved September 15, 2026. Profound publishes only the monthly equivalent of a 12-month commitment. Peec shows the plans and the full feature table but no figure at all in the served page.
Peec AI took its pricing offline. The plans page lists four tiers, the full feature matrix and the annual discount, and no numbers. The prices render through JavaScript and never reach the HTML. It is an AI visibility vendor whose own price a language model cannot read.
Which engines does each tool cover?
Engine coverage is where list price misleads most. Two tools at the same price can differ by four surfaces.
| Tool | In the reference plan | Claude | Maximum possible |
|---|---|---|---|
| Otterly.AI | ChatGPT, AI Overviews, Perplexity, Copilot | US$109 add-on on Standard | 7 with add-ons |
| Scrunch | ChatGPT, Claude, Gemini, Perplexity, AI Mode, AI Overviews, Meta | included | 7 |
| Profound | ChatGPT, Perplexity, AI Overviews | Enterprise only | 9 on Enterprise |
| Peec AI | 3 chosen from 6 surfaces | Enterprise only | 12 on Enterprise |
| Semrush | not verifiable | not verifiable | not verifiable |
| Ahrefs Brand Radar | AI Overviews, AI Mode, ChatGPT, Perplexity, Gemini, Copilot | in custom prompts | 7 |
| SE Ranking | AI Overviews, AI Mode, ChatGPT, Gemini, Perplexity | no | 5 |
Two inconsistencies on record. Scrunch lists 7 engines on its pricing page, 5 on its product page and a different set of 7 names on its insights page. Semrush names 4 engines on pricing, 3 on features and only 2 in the official product documentation, which is why that row reads not verifiable rather than a number.
Which ones work outside English?
This criterion eliminates more candidates than any other, and almost no English-language roundup tests it.
| Tool | Brazil | Portuguese | Where it is documented |
|---|---|---|---|
| Otterly.AI | yes | yes | Help center, named list of 68 countries; monitoring declared for any language |
| Semrush | yes | yes | Product documentation: Brazil in regional coverage, language selected at setup |
| Scrunch | not verifiable | not verifiable | Only the generic phrase “regions and languages”, with no list |
| Profound | not verifiable | not verifiable | The page counts how many languages and regions are included without naming any |
| Peec AI | not verifiable | not verifiable | Claims “any supported country or language” without publishing which |
| Ahrefs Brand Radar | not verifiable | not verifiable | No country or language claim at all |
| SE Ranking | not verifiable | not verifiable | Runs a Portuguese website, which is interface language, not tracking coverage |
A distinction that trips up many buyers: a vendor site translated into your language is not evidence of tracking coverage in that language. They are different things and only the second one matters.
Not verifiable does not mean it does not work. It means the vendor will not commit in writing, and that the risk of finding out on the second invoice is yours. For any of those five, the honest path is a paid 30-day test with real prompts in the target language before recommending it to a client.
Which one works for multiple clients?
Consultancies and agencies need several brands under one subscription. The market splits cleanly in two here.
| Tool | Brands or projects in the reference plan | Works for an agency? |
|---|---|---|
| Otterly.AI | unlimited workspaces on Standard and Premium | yes, with an agency program and bonus prompts |
| Peec AI | 2 projects on Pro, 5 on Advanced | with a cap |
| Scrunch | not stated on the pricing page | not verifiable |
| Profound | 1 brand on Starter and Growth | no, outside Enterprise |
| Semrush | not verifiable | not verifiable |
| Ahrefs Brand Radar | not verifiable | not verifiable |
| SE Ranking | not verifiable | not verifiable |
The single-brand cap outside Enterprise is what removes Profound from any multi-client operation, however good the platform is.
Does the data leave the platform?
A client report is not assembled inside the vendor's dashboard. A tool that only ships an interface forces manual export, and manual export becomes a transcription error within three months.
| Tool | API | MCP | BI connector |
|---|---|---|---|
| Otterly.AI | yes, 2,000 requests per month on Standard | yes, same allowance | Looker Studio in the agency program |
| Scrunch | Enterprise only | yes, on every plan, read and write | not verifiable |
| Peec AI | yes | yes | community Looker Studio connector |
| Profound | Enterprise only | not verifiable | not verifiable |
| Ahrefs Brand Radar | yes, at platform level | yes, at platform level | not verifiable |
| Semrush | not verifiable for this product | not verifiable | not verifiable |
| SE Ranking | not verifiable | not verifiable | not verifiable |
MCP matters more than it looks: with it, an assistant reads the series straight from the tool and builds the report cut with no intermediate export. Scrunch opened MCP to every plan in May 2026, which reversed an advantage that used to be Otterly's.
What does it cost per client?
List price is not the number that matters to anyone serving a book of clients. The number that matters is the per-client split. The table below assumes 15 monitored prompts per client, which is our calibration for a mid-sized asset manager.
| Tool | Plan | Clients that fit | US$ per client/month |
|---|---|---|---|
| SE Ranking | Core + AI Search | 30 | US$7 |
| Scrunch | Growth | 46 | US$11 |
| Scrunch | Starter | 23 | US$13 |
| Otterly.AI | Premium | 26 | US$19 |
| Otterly.AI | Standard | 6 | US$32 |
| Semrush | Pro+ | 6 | US$50 |
| Profound | Growth | 1 | US$399 |
Our calculation, not the vendor's: plan prompt volume divided by 15, rounded down, with the month-to-month price divided by the result. Where a vendor caps the number of brands, the brand cap wins, which is what puts Profound at US$399. Ahrefs is excluded because it bills per check, which does not convert into prompts on a comparable basis.
What changed since July 2026?
We ran this same evaluation in July 2026. Re-running it in September produced the most useful finding of the whole comparison: in eight weeks, five of the seven vendors changed something that would have invalidated a recommendation.
| Vendor | Recorded in July | Verified on September 15, 2026 |
|---|---|---|
| Ahrefs Brand Radar | US$328 per month, no Claude | US$179 for the equivalent arrangement, and Claude covered in custom prompts |
| Peec AI | ChatGPT, Perplexity and Gemini only | 6 surfaces in the pool, 12 on Enterprise, and pricing no longer public |
| Scrunch | scrunchai.com domain, no MCP | scrunch.com domain, MCP on every plan since May |
| Semrush | 4 engines | no official source publishes a closed list; not verifiable |
| Otterly.AI | confirmed | confirmed, no pricing discrepancy |
Ahrefs is the extreme case: a price 45% above the real one would have gone into a public piece. The July comparison was finished and unpublished; shipping it without re-verification would have cost exactly the credibility this material exists to build.
The operational conclusion is not about any vendor. It is about shelf life: a tool figure has a half-life measured in weeks. Any comparison without a date beside the number, this one included, should be treated as stale by default.
Which one should you pick?
There is no best tool. There is the one that survives your constraints.
| Dominant constraint | Pick | Why |
|---|---|---|
| Non-English operation | Otterly.AI or Semrush | The only two with publicly documented country and language coverage |
| Agency with many clients | Otterly.AI Premium | Unlimited workspaces, an agency program and the best split among the language-documented options |
| Lowest cost per prompt | SE Ranking | US$7 per client is an order of magnitude below the rest, with no documented language coverage |
| Most engines without add-ons | Scrunch | Seven surfaces included and MCP on every plan |
| Single brand, large budget | Profound | The single-brand cap stops being a problem and the platform is mature |
| Already an Ahrefs or Semrush customer | your own vendor's add-on | Low marginal cost and one fewer account |
Our standing recommendation: no final choice without a paid 30-day test using the client's real prompts, on the same days, against your own collection. The divergence between instruments will exist, and that divergence is itself primary data.
And the caveat that costs us money to say. A market tool does not hand back the auditable raw output of each run, and its definition of “cited” is a vendor black box. When that definition changes, the client series breaks without warning. That is why we keep our own collection alongside the tool, and why the monthly report states which instrument produced each number. Anyone reporting only the vendor dashboard is reselling the vendor dashboard.
Frequently asked questions
Which AI visibility monitoring tool works in Portuguese?
Otterly.AI and Semrush are the only two that publicly document coverage for Brazil and language selection, verified on September 15, 2026. Otterly publishes a named list of 68 countries including Brazil and states that monitoring works in any language. Semrush lists Brazil in its regional coverage and lets you choose the language at setup. The other five platforms publish nothing verifiable on the subject.
Can one subscription cover several clients?
It depends on the vendor, and the gap is wide. Otterly.AI offers unlimited workspaces on Standard and Premium. Peec AI caps the Pro plan at 2 projects. Profound caps Starter and Growth at a single brand, which rules it out for agency use. Scrunch, Semrush, Ahrefs and SE Ranking do not state the limit on their public pages.
What does AI visibility monitoring cost per client?
Between US$7 and US$399 per client per month, assuming 15 monitored prompts per client. The cheap end is SE Ranking at US$7. The expensive end is Profound at US$399, because the plan covers one brand and no split is possible. Among the tools with documented language coverage, the best arithmetic is Otterly.AI Premium at roughly US$19 per client.
Does a monitoring tool replace your own collection?
It does not, for three reasons. No platform returns the auditable raw output of each run, which is what lets you re-judge without re-collecting. The definition of “cited” belongs to the vendor and can change without notice, which breaks the historical series silently. And the series is hostage to the contract: if the tool changes methodology, raises prices or shuts down, the client history goes with it.
How often do these prices change?
Faster than any published material can keep up with. Between July and September 2026, five of the seven vendors on this list changed price, packaging, engine coverage or domain. One of them dropped roughly 45% on the equivalent arrangement. No tool figure belongs in a commercial proposal without the date of the lookup beside it.
ABOUT THIS ANALYSIS
Research conducted by Julio Bozzi and Felipe Reusch, founding partners of ReBo AI Visibility. Julio is COO, an economist trained at Insper with over 10 years in fintech and private markets. Felipe is CEO, with over 10 years across asset management, technology and AI. All data was read directly from each vendor's official pricing pages and documentation on September 15, 2026. ReBo holds no affiliate, reseller or commission agreement with any of the platforms listed. Corrections: contact@rebovisibility.com.