How-to · 3 steps
How to check whether AI answers cite you
with Monid.
Give this to your agent
$set up https://monid.ai/SKILL.md, then tell me which brands the AI answers cite for <question>
and let it take it from there.
#QuestionWho got citedCost
1Scraping APIsFirecrawl, Bright Data, Apify$0.0204 ✓
2Tool callingNango, Composio, Arcade$0.0204 ✓
3Our own pitchtwo marketplaces, by name$0.0156 ✓
4Zapier rivalsn8n, Composio, Make$0.0204 ✓
5The cheap enginesame answer, 66x less$0.00015 ✓
6Timed out504 twice, billed nothing$0.00 ✓
7Uscited in none of the eleven$0.00 ✓
8Totaltwelve answers, two free fails$0.0770 ✓
four questions · three engines · eleven answers · zero mentionsagent running$0.0770
Real run, 2026-09-24. Competitors named only where an answer named them.
Step 1
Set up Monid.
One line. It installs the CLI and asks for an API key from app.monid.ai. New accounts start with $1.00.
Say this to your agent
>set up https://monid.ai/SKILL.md
Step 2
Pick what you need. Say it in one sentence.
Click a job. Paste the sentence, fill in the brackets.
What Google's AI answer says
Say this to your agent
>ask google ai mode <question> with litescrape on monid and give me the answer plus every source it cites, in order. check its health flag first and budget three attempts on a 503, not one: on our second run it took two failures before a 200
What the agent runsdone
litescrape /google/ai-mode$0.00015
litescrape /google/ai-overview · not run$0.00015
mrscraper /google/ai-mode$0.01
cloro /google/ai-mode · not run$0.008
mrscraper /gpt/web-search · not run$0.025
What came backfour buyer questions · 2026-09-24
Same question, same surfaceprice per answer
what came backthe generated answer plus ten cited references, in the order the answer used them
the price$0.00015 against $0.01 for the same surface, which is 66x
the catchit was flagged degraded on both runs and 503'd on both; once a single retry cleared it and once it took three attempts, so one retry is not a budget
a 503 still reads as COMPLETEDthe run record shows a completed status with the 503 inside it, so a script filtering on status ingests the failure silently; read the provider's own status code
the answer variesanswers are generated per request, so two calls the same day differ
$0.00015per question
What ChatGPT says, and what it links
Say this to your agent
>ask chatgpt <question> with cloro on monid, passing a country: it is required on every engine this provider fronts and none of these sentences used to supply it. give me the cited sources as domains, in order
What the agent runsdone
cloro /chatgpt/ask$0.008
cloro /copilot/ask · not run$0.008
cloro /gemini/ask · not run$0.008
mrscraper /gemini/ask · not run$0.01
What came backfour buyer questions · 2026-09-24
the shapethree to five cited domains per answer, which is tighter than the other engines
on our own pitchasked the question this company's homepage answers, it named two smaller API marketplaces instead
the metaphoron the tool-calling question it cited the product we compare ourselves to, not us
the only engine with a brand fieldthis one returns an entities block naming the brands; the other four return free text you have to read yourself, which is why a brand-level metric cannot be built uniformly across them
the reada tight citation list means fewer slots, so being absent costs more here than on a ten-source answer
$0.0056per answer
What Perplexity says, when it answers
Say this to your agent
>ask perplexity <question> with cloro on monid, with a country set. read its health flag at run time rather than trusting this page: it was degraded and timing out on our run and stable and answering in forty seconds on the next one. if it does time out twice, record the miss rather than retrying again
What the agent runsdone
cloro /perplexity/ask$0.008
What came backfour buyer questions · 2026-09-24
the shapeten sources per answer, the widest citation list of the three engines
the failureone of four calls timed out twice with a 504 and billed nothing either time
the flag movesit was degraded when this page was built and stable when the page was cold-tested, answering four times out of four, so treat any health note here as a snapshot and check it yourself. The flag on the other engine did not move
what to dotreat a missing engine as missing data, not as a zero mention
$0.0048per answer
A rate, not a single answer
Say this to your agent
>put these four buyer questions to every answer engine on monid, each with a country set, and give me TWO tables: one counting answers that name my brand in the text, one counting answers whose cited sources include my domain. tell me both numbers and do not average them
What the agent runsdone
litescrape /google/ai-mode$0.00015
cloro /chatgpt/ask$0.008
cloro /perplexity/ask$0.008
mrscraper /google/ai-mode$0.01
What came backfour buyer questions · 2026-09-24
Cited across eleven answersshare of answers
the answereleven answers came back and cited this company in none of them
there are two shares, not oneon a second category the same twenty answers gave a brand named in the text on twenty of twenty and the domain in the sources on three of twenty; both are defensible and they are not the same measurement
and they invert the rankinga rival named in one answer of twenty held the highest domain-citation share, above the brand that was named in all twenty, so choosing the other metric tells the opposite story
why four questionsone answer is a draw; four questions across three engines is a rate you can track
what it costunder nine cents to find out you are invisible in your own category
the engines are not standing in the same placeone provider pins the query to a country and the other sends no location at all, so the engines in your table are not asking from the same country unless you set it everywhere
the one blankone AI Mode answer named no vendor at all, so nobody could have been cited on it
$0.0770four questions, three engines
Who got cited instead
Say this to your agent
>from those answers, rank the brands twice: once by how many answers name them in the text, once by how many cite their domain. show me both columns side by side and tell me where the two orders disagree
What the agent runsdone
cloro /chatgpt/ask$0.008
cloro /perplexity/ask$0.008
What came backfour buyer questions · 2026-09-24
the leaderone competitor appeared in more than a third of the answers without ever being asked about by name
the closest callon the question nearest our own pitch, two much smaller marketplaces were named and we were not
per enginethe leaders differ by engine, so a single-engine check would have given the wrong picture
the two columns disagreeon a second category the brand named most often and the brand cited most often were different brands, so a single-column ranking is a choice you have to declare
and the sources may not match the answerone engine's cited domains had no overlap at all with the brands its own answer recommended, so counting citations is not counting recommendations
no extra callthis is counting, not calling: the answers were already paid for
$0.00already in the answers
How often AI answers cite a domain
Say this to your agent
>get the ai responses count for <domain> with ahrefs on monid. it is $0.36 a call, so confirm with me before you run it
What the agent runsdone
ahrefs /site-explorer/ai-responses-count · not run$0.36
What it returnsshape of the answer
what it isa count of how often AI answers referenced a domain, as one aggregate number
the price$0.36, which is more than this whole four-question sweep cost
when it earns itfor tracking one number over months, not for finding out who gets cited today
it was run, and it disagreeson a second domain it reported a zero for one engine that another engine cites over thirteen thousand times, so a zero here is missing coverage rather than absence. Also worth knowing: the sweep above answered the question for eight cents, so this one stayed in the catalog
$0.36per domain · catalog rate
Step 3
Take the cheapest route with Monid.
Ask the engine you can get for a hundredth of a cent first. Pay for the ones you cannot, and only then for a number that aggregates them.
Cheap engine · $0.00015
AI Mode$0.00015answer plus references
AI Overview$0.00015the other Google surface
The engines you must pay for
Perplexity$0.0048
ChatGPT$0.0056
Copilot$0.008
MrScraper AI Mode$0.01
GPT web search$0.025
Count it
Tally the brands$0.00
AI responses count$0.36one aggregate number · catalog rate
Back to you
rowanswered bycost
1this engine named none$0.0001
2ChatGPT$0.0057
3Perplexity$0.0049
4MrScraper AI Mode$0.0238
5ChatGPT$0.0057
5 of 5 rows$0.0403
Say this to your agent
>put <question> to every answer engine on monid, passing a country, which one provider requires on every call. do not stop when an engine names nobody: on our run one engine named none while another named two, so sweep them all and tell me which brands each one cited and what every answer cost
“cheapest engine first”the same Google surface is $0.00015 through one provider and $0.01 through another
“ask four questions”one answer is a draw; a rate is something you can watch move
“say which share you mean”named in the text and cited as a source gave 100% and 15% on the same twenty answers, and ranked the brands in opposite orders
“count, do not re-ask”ranking the brands is arithmetic on answers you already bought
One key. 1,700+ tools.
AI answers are one aisle.
AI Mode and AI Overviewlitescrape · $0.00015 / call
ChatGPT, Perplexity, Copilotcloro · $0.008 / answer
AI Mode, Gemini, GPT searchmrscraper · $0.01 to $0.025
How often AI answers cite youahrefs · $0.36 / call
+ 1,700 more across 55 providers
Browse the catalog →Read more
View all →
Claude for SEO and GEO
One loop from keyword to answer engine, and what the whole thing cost.
Bing Search API retired: the alternatives
What replaced it, what each one charges a query, and which an agent should call.
Extract page content for RAG
Once you know which pages an answer cites, reading them is the next call.