AEO tracking tools for Hypertune

For Som to hand to Vinny. Prices checked 18 October 2026 from the vendor pages named below. This is a one-month trial choice, not a longer plan.

Already decided, and not reopened here: AEO stays in-house, on top of the existing SEO calendar (two pages a week). Bobby's 25-question list is not used, and it is not the basis of the hours. The volume below is about 25 questions of our own, United States only, once a week, on ChatGPT, Perplexity, Gemini, Copilot, and Google AI Overviews.

What to trial

Trial Ahrefs custom prompts for one month, the $50 pack (2,500 checks), weekly, one US location, those five engines.

Run a free script beside it on the same 25 questions, limited to ChatGPT and Perplexity. At the end of the month, keep one.

Leave the Brand Radar index ($199 per platform, $699 for all) off this trial. That product is a huge pre-built prompt database. Ahrefs says it is thin for brands with little search volume. Hypertune's job this month is a fixed list of buyer questions, which is the custom-prompt product.

Prices, and what each tier tracks

Checked 18 October 2026. All figures are USD, before tax, as shown that day. These pages change. Re-check at checkout.

Ahrefs Brand Radar

Two products share the Brand Radar name. The quote Vinny already has ($199 for one AI platform index, $699 for all tools, and $50 for 2,500 prompt checks) matches the public pages.

AI Visibility Index. Starts at $199 a month per AI platform, or $699 a month for all models. Sources: ahrefs.com/brand-radar and the standalone-tools section of ahrefs.com/blog/ahrefs-pricing/.

What it tracks: a pre-collected set of search-backed prompts. The product page said 485 million or more on 18 October. Platforms in the index: Google AI Overviews, Google AI Mode, ChatGPT, Perplexity, Gemini, Copilot. Metrics: brand mentions in the answer text, citations (links to a page or domain), share of voice against competitors, and estimated impressions weighted by search volume. Question sets are retested monthly. No prompt setup. History back to 2025 is claimed on the product page. YouTube and Reddit are separate indexes on the same product, not AI answers.

Ahrefs' own FAQ says this index is built from their keyword database and can have limited coverage when a brand has little search demand. Custom prompts are the path they describe for a brand that is not well known, or for uncommon questions.

Buying the full index also includes 2,500 custom-prompt checks a month. You do not need the index to buy those checks on their own.

Custom prompts. The $50 add-on. From the pricing article and the help article How to set up custom prompts (help article dated 3 September 2026, read 18 October):

Pack	Checks a month	Price	Overage per check
Basic	2,500	$50	$0.020
Growth	7,000	$100	$0.015
Scale	25,000	$250	$0.010

One check is one prompt, one engine, one location, one update. Engines: AI Overviews, AI Mode, ChatGPT, Perplexity, Gemini, Copilot, and Claude. Claude uses 8 checks per update, so leave it off. Grok is listed, with a note that Ahrefs cannot collect new Grok data for now. Cadence is daily, weekly, or monthly, per engine. First answers can take up to 24 hours. The pricing article says custom prompts can be bought without a base Ahrefs plan. Confirm that at checkout. If the cart forces a base plan, stop and tell Vinny before paying.

What Lite includes. Lite is $129 a month and includes 150 custom-prompt checks. Standard ($249) includes 300. Advanced ($449) includes 600. Starter ($29) includes none. The help article calls the included checks a built-in trial of AI visibility tracking. It is not the index, and it is not the $50 pack. That matches the decision already made.

Check math for this list, weekly, US, five engines: 25 prompts x 5 engines x 4 weeks = 500 checks. The $50 pack covers that about five times over, so there is room for a second pass or a few extra questions. Daily on the same list is 25 x 5 x 30 = 3,750 checks, which spills past 2,500 (overage about $25, or the $100 pack). Run it weekly.

Lite's 150 checks cover about 7 prompts on all five engines each week, or the full 25 prompts on one engine. Not this job.

Peec

Read on peec.ai/pricing, monthly prices on the cards, 18 October 2026.

Plan	Monthly price	Prompts	Models	Projects	Cadence
Starter	$95	50	choose 3	1	daily
Pro	$245	150	choose 3	2	daily
Advanced	$495	350	choose 3	5	daily
Enterprise	custom	custom	up to 13	unlimited	daily or weekly

Unlimited users on every plan. The page's own example of "AI answers" is prompts x models x days, so Starter is 50 x 3 x 30 = 4,500 answers a month. Self-serve plans are daily only. Weekly is an Enterprise option. The FAQ on the page says annual billing is 15% off. I did not toggle the cards, so I am not quoting an annual dollar price.

Models listed for Starter, Pro, and Advanced in the comparison table: ChatGPT, AI Mode, AI Overviews, Microsoft Copilot, Gemini, Naver AI. Perplexity is on the Enterprise row and in the additional-models list. Peec's comparison page peec.ai/peec-vs-profound says you can pick Perplexity as one of the three on every base plan. Those two Peec pages disagree. Confirm Perplexity at signup before counting it as included.

Extra-model prices shown lower on the same page: $30 (Starter), $70 (Pro), $140 (Advanced), with "Save $60", "Save $180", and "Save $300" beside them. I could not tell whether those three figures are the monthly rate or the annual rate. Do not budget a five-engine Peec total from this note without looking at the cart.

Five engines means three included plus two extras, on top of $95, and you pay for a daily run when this team wants a weekly one.

Profound

Read on tryprofound.com/pricing, 18 October 2026.

Trial: free for 7 days. 50 prompts, run daily, on ChatGPT, Gemini, and Google AI Overviews. The prompt list is Profound's, and the trial FAQ says you cannot change it. No exports, no API, one language, one region.
Enterprise: custom quote. Your own prompts, daily, up to 9 answer engines, exports, API, SSO. No public monthly price for a brand.

There is no self-serve plan to buy for a month and cancel. Older public prices ($99 and $399) are not on this page. Skip Profound for this trial.

Otterly

Read on otterly.ai/pricing, monthly tab, 18 October 2026.

Plan	Monthly price	Prompts	Engines included	Cadence
Lite	$29	15	ChatGPT, AI Overviews, Perplexity, Copilot	daily
Standard	$189	100	same four	daily
Premium	$489	400	same four	daily
Enterprise	custom	from 1,000	same, plus custom terms	daily

Gemini, Google AI Mode, and Claude are add-ons on every plan. Monthly add-on prices on the page:

Add-on	Lite	Standard	Premium
Gemini	$9	$59	$149
Google AI Mode	$9	$59	$149
Claude	$29	$109	$439

Extra prompts: 100 more for $99 a month on Standard or Premium. Not available on Lite. Standard can add up to 300 extra prompts, then Otterly moves you to Premium.

The page offers annual billing at 15% off. I read the monthly tab only.

For this list: Lite plus Gemini is $38 and stops at 15 prompts. Standard plus Gemini is $248 and tracks 100 prompts every day. Both are a worse fit than $50 for 25 questions once a week on five engines.

Free or open-source method

OneGlanse (MIT). README at github.com/oneglanse/oneglanse, read 18 October 2026. It runs your prompts in the real product websites for ChatGPT, Perplexity, Gemini, Claude, and Google AI Overview, then scores the saved answers with an API key you supply. Copilot is not in that list. Software cost is zero. You still need accounts on those products, Docker, Node, and an analysis-model key. A server schedule needs a residential proxy. Browser runs break when a site changes its login or blocks the browser. Treat install as a day, not an afternoon, and expect a broken week.

A small script, which is the free method to run beside the trial. Same 25 prompts, ChatGPT (with web search) and Perplexity only, written into the same sheet Hoa uses. OneGlanse's README cites a Surfer study of 1,000 prompts in which brand names from the product website and from a comparable API overlapped 21.3% to 31.6%. The script is a check on whether the paid tool matches a direct query. It is not a second full measurement of all five engines. Hoa opens the real ChatGPT and Perplexity sites for about five prompts a week and notes whether the script saw the same brand and the same citation.

Do not script Copilot, Gemini, and AI Overviews in this month. That is where the setup time goes, and the paid tool already covers them.

Hours

125 answers a week (25 questions x 5 engines). Setup is once. Weekly is every week after that. Ranges are judgments from the volume, not a timed pilot. Bobby's questions are not in these numbers.

Way of checking	Setup	Weekly, after setup
Manual (open each site, paste, log a short row)	4 to 6 hours	4 to 6 hours
Script or skill (ChatGPT and Perplexity only)	8 to 14 hours	1.5 to 3 hours when it runs; about 4 when it breaks
Paid tool (Ahrefs custom prompts, weekly)	2 to 4 hours	1.5 to 3 hours

Manual weekly assumes about 1.5 to 2 minutes per answer to paste, wait, and log mention, citation, and competitor, then a short pass on what changed. Reading every answer in full is longer, toward 6 to 8 hours. Hoa's 2 to 3 hours does not cover a manual week.

Script setup is a small script and a sheet, not OneGlanse. OneGlanse, if you later want the real sites, is another 1 to 2 days, and Copilot is still missing.

Paid weekly split: Hoa about 1 to 2 hours copying what changed into the log, inside the 2 to 3 hours she can spend. Som about 1 to 1.5 hours deciding whether either of the two pages that week should answer a miss. Austin only if a new claim needs a signature, which is already how a page ships.

Extra time on the SEO calendar

Closer to 20% extra time.

Som's timesheet has the SEO calendar at about 20 hours a week once November settles, with two pages a week as the core. On that base, 20% is about 4 hours. That matches the paid-tool week: Som's 1 to 1.5 hours on the log, plus about an hour a page to write a direct answer an engine can quote, on the two pages already being drafted.

What drives it:

Rewriting passages is the steady extra. It sits inside the two pages. It is a clearer answer, a citable fact, and a source. It is not a third page.
Tracking stays inside Hoa's 2 to 3 hours if the $50 tool works. Manual tracking is a separate half-job and would push the week well past 20%.
Off-site work (Reddit, YouTube, review sites, outreach) is not in the calendar. Starting it makes this 50% or more. Leave it out of the trial month.

Uncertainty: this is not a timed pilot. Some weeks the log will show a wrong claim on a page that is already live (a safety line, an FPS figure, a price). Fixing that, and getting Austin's sign-off, can make that week look closer to 40%. If most weeks are "nothing to change on the two drafts," the extra can land under 20%. I would not call it 50% unless someone adds off-site work or keeps checking by hand.

Rank, for this team

Som owns the prompt list and the weekly read. Hoa has about 2 to 3 hours for the log. Austin signs claims and does not run the tool.

Ahrefs custom prompts, $50, one month. Weekly, US, five engines, about 25 prompts. Fits the check math, the hour budget, and an account the team can cancel. This is the trial.
Free script on ChatGPT and Perplexity, same 25 prompts, run in the same month. Hoa spot-checks five prompts in the real sites. This is the side-by-side, not a second paid tool.
Otterly Lite plus Gemini ($38) only if the Ahrefs $50 cart fails. It covers the five engines and stops at 15 prompts, with no way to add prompts on Lite. Standard plus Gemini ($248) covers 100 prompts a day, which is more tracking than this list needs.
Peec Starter ($95). Three models, daily, and a checkout question on whether Perplexity is included. Five engines needs two add-ons at a price I could not pin down on 18 October. Skip for the trial.

Profound stays off the list. There is no public price to approve, and the free trial will not run our prompts.

After the month, keep the paid tool if Hoa stayed inside 2 to 3 hours and the script usually agreed with it on ChatGPT and Perplexity. Drop the paid tool if the script was enough and the breakage was rare. Do not keep both.