GEO

How GEO Tools Count Prompts: Once per Prompt, per Engine or in Credits

On this page
  1. What the ten tools do, at a glance
  2. Which tools count a prompt once, however many engines answer it?
  3. Does the country count as well as the prompt?
  4. Which tools count every engine run?
  5. Which tools do not say how they count?
  6. Why does the counting rule change what a plan costs?
  7. Where I got these facts

Most GEO tools count a prompt once however many AI engines answer it, a few count every engine run, and some meter usage in credits. The difference decides what a plan really covers: 50 prompts can be 50 units of usage on one tool and 400 on another. This page sets out the rule for ten tools, in the vendor’s own words where I have them, with the source and the date of each fact.

Disclosure. This site has affiliate links for LLM Pulse, Otterly, RankScale and Trakkr elsewhere. Every link on this page is a plain vendor URL, the table is in alphabetical order, and no tool is ranked or recommended here. I have not tested the counting of any tool myself: every rule below is a vendor statement or a vendor page, recorded with its date.

What the ten tools do, at a glance

How each tool counts usage (alphabetical; source and date for every row in the sections below)
ToolWhat is countedSource and dateVendor-confirmed
KollavoNot disclosedVendor email 31 Oct 2026; pricing page read 2 Nov 2026Not stated by the vendor
LLM PulsePrompts per plan, with five AI models included; the rule across models is not statedPricing page read by Marco, 19 Oct 2026; vendor email 10 Nov 2026Plan row confirmed; counting rule not stated
MentionXOnce per saved prompt, however many engines answerVendor email, 7 Nov 2026Yes
OneLenceOnce per question, however many engines answerVendor email, 15 Nov 2026Yes
OtterlyOnce per prompt across the four default engines, counted per countryVendor email, 19 Oct 2026Yes
PeecOnce per prompt, however many models run itVendor email, 26 Oct 2026Yes
PinepromptPrompts count once; a second monthly limit counts each run on each engine as a responseVendor email, 3 Nov 2026Yes
RankScaleCredits, charged per engine runVendor emails, 19 Oct and 11 Nov 2026Yes
RankScopeOnce per prompt, however many engines run itVendor email, 14 Nov 2026Yes
TrakkrOnce per prompt, which runs on all eight modelsVendor email, 31 Oct 2026Yes

“Vendor-confirmed” means the vendor wrote the rule to me in an email on that date. It does not mean I checked it in the product, and a vendor can change a rule after the date shown. The GEO tool pricing index keeps the same dated record and the unit each vendor uses, and never converts one unit into another.

Which tools count a prompt once, however many engines answer it?

MentionX, OneLence, Peec, Pineprompt, RankScope and Trakkr all say a prompt counts once against the plan limit. Otterly follows the same idea with one extra variable, the country.

  • MentionX. One saved question counts as one prompt regardless of how many engines answer it. Its team gave an example in its email of 7 November 2026: Spark’s 100 prompts across its three included engines count as 100 prompts, not 300. The vendor also wrote that Google keyword tracking is included and does not use up the prompt allowance; that part is vendor-written and I have not verified it in the product.
  • OneLence. One question counts once, however many engines answer it (vendor email, 15 November 2026). AI answer tracking is an add-on pack with up to 25 active questions per website, checked weekly plus one manual run per website per day.
  • Peec. I asked whether one prompt tracked on three models uses one prompt or three. The reply of 26 October 2026 was “It counts as 1 Prompt!”, so a prompt is counted once however many models run it.
  • Pineprompt. A prompt counts once against the prompt limit however many engines it runs on (vendor email, 3 November 2026). There is a second limit, covered in its own section below.
  • RankScope. A prompt counts once however many engines it runs on (vendor email, 14 November 2026). Every tracked prompt is scanned once a day on all plans.
  • Trakkr. One prompt counts once against the limit and runs on all eight models. The founder’s email of 31 October 2026 put it as “50 prompts is 400 answers a day, or 12,000 a month”.

Does the country count as well as the prompt?

At Otterly, yes: a prompt is counted per country. Its team wrote on 19 October 2026 that one prompt is monitored across the four default engines (ChatGPT, Google AI Overviews, Perplexity and Copilot), that a prompt is counted per country, and that countries are not capped while prompts are. Tracking the same prompt in two countries therefore uses two prompts. Pineprompt has a similar unit in a different place: a monitor is one country plus one language, and its plans cap the number of monitors (vendor email, 3 November 2026).

Which tools count every engine run?

RankScale meters credits, and each run on each engine is charged separately. Its team wrote on 19 October 2026 that ChatGPT, Perplexity, Google AI Overviews, Google AI Mode and Gemini cost 0.25 credits per run each, and that model engines such as Claude, DeepSeek and Mistral cost 1 to 4 credits per run. On 11 November 2026 it wrote that Microsoft Copilot stopped being tracked on 8 November 2026 and that Bing AI Mode replaced it at the same 0.25 credits per run. The same prompt on more engines, or on a more frequent schedule, uses more credits.

Pineprompt also counts engine runs, but as a second limit. Each run on one engine is one response, and each plan has a monthly response allowance (Starter 6,000, Professional 18,000 and Business 42,000, vendor email of 3 November 2026). Both limits are in the vendor’s plan row, so a plan can run out of responses before it runs out of prompts. That is my reading of the two limits, not a statement from the vendor.

Which tools do not say how they count?

Kollavo has not stated a prompt count, and LLM Pulse has not stated the rule across engines. Kollavo’s written terms of 31 October 2026 do not disclose a prompt count, and its pricing page read on 2 November 2026 does not state one either, so I cannot say how it counts. LLM Pulse lists prompts per plan in euros (Starter 50, Growth 150, Scale 450, read by Marco on 19 October 2026) with five AI models included on every plan, which the vendor confirmed on 10 November 2026 (ChatGPT, Perplexity, Gemini, Google AI Mode and Google AI Overviews). Extra models are paid add-ons. Whether a prompt on five models counts once or five times is not in my record, so I do not claim either.

Why does the counting rule change what a plan costs?

Because the same plan label hides different amounts of work. A prompt limit says how many questions you can track. It does not say how many answers the tool collects, how often, or in which countries. Three things are worth asking a vendor before you buy:

  • What is one unit? A prompt, a question, a credit, a response or a monitor.
  • Does a second engine or a second country cost another unit? For Peec, MentionX, OneLence, RankScope and Trakkr the vendor says no for engines. For Otterly a second country does, and for RankScale a second engine does.
  • Is there a second limit? Pineprompt’s monthly responses are one example.

Prices and rules change, so check the vendor’s own pricing page before you buy. For plans compared by your own prompt count and engines, use the GEO plan matcher, and for a wider comparison see the best GEO tools. If two of these names look alike, see RankScope vs RankScale.

Where I got these facts

Every rule above comes from a vendor email to me or a vendor pricing page, with the date given in the table: MentionX, OneLence, Otterly, Peec, Pineprompt, RankScale, RankScope, Trakkr, Kollavo and LLM Pulse. None of the data was tested by this site, and the quality of each tool’s answers is not compared here.

Frequently asked questions

Does a GEO prompt count once per AI engine?

It depends on the tool. MentionX, OneLence, Peec, Pineprompt, RankScope and Trakkr each said a prompt counts once however many engines answer it (vendor emails, 19 October to 15 November 2026). RankScale charges credits per engine run (its team, 19 October 2026).

What is the difference between a prompt and a response?

A prompt is the question you track. A response is one answer from one engine. Pineprompt counts both: prompts against a prompt limit and each engine run as a response against a monthly allowance (vendor email, 3 November 2026).

Are prompts counted per country?

At Otterly, yes. Its team wrote on 19 October 2026 that a prompt is counted per country and that countries are not capped while prompts are.

How do credits work at RankScale?

Every run of a prompt on each selected engine is charged separately. ChatGPT, Perplexity, Google AI Overviews, Google AI Mode, Gemini and Bing AI Mode cost 0.25 credits per run, and model engines such as Claude, DeepSeek and Mistral cost 1 to 4 credits (its team, 19 October and 11 November 2026).

Which tools do not state a counting rule?

Kollavo has not disclosed a prompt count (31 October and 2 November 2026). LLM Pulse lists prompts per plan with five models included but has not stated whether a prompt counts once across them (10 November 2026).

Published Last updated

← All posts

↑↓ navigate↵ openesc close