How-to · 3 steps
How to read a dealer's lot
with Monid.
Give this to your agent
$set up https://monid.ai/SKILL.md, then show me which units on this lot are priced wrong and why
and let it take it from there.
#JobWhat came backCost
1The radius10,382 within 25 miles$0.01 ✓
2A rival's lot29 vehicles, five brands$0.01 ✓
3The value$39,530 vs a $36,995 ask$0.01 ✓
4The rating4.8 out of 5 for the year$0.01 ✓
5Auction comps$31,250 · $36,750 · $23,900$0.01 ✓
6What one sold for$30,250 at auction$0.01 ✓
7The dealer3.9 of 5 across 21 reviews$0.01 ✓
8The dead end165 listings, every id null$0.06 ✓
9Totaltwelve calls, one free fail$0.16 ✓
a rival's whole lot · one unit priced three ways · one dealer checkedagent running$0.16
Real run, 2026-09-24. Businesses named, no reviewer named.
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.
A rival's entire lot
Say this to your agent
>pull the inventory for <dealer id> from cars.com on monid. the id is a uuid you only get from a $0.03 search first, the response carries no total and no page count so paging runs until a page comes back empty and each empty page still bills, and new and used stock are mixed with no filter, so split them yourself before saying what the lot leans toward
What the agent runsdone
cars.com /get_dealer_inventory$0.01
cars.com /search_listings$0.03
edmunds /search_inventory · not run$0.01
What came backCharlotte NC · 2026-09-24
the lot29 vehicles in one call, from $20,914 to $57,968
the mixfive brands, new and used stock side by side on the same lot
there is no end signalthe response has no total and no page count, so paging is blind: on a second dealer pages twenty and fifty each came back empty and each billed a cent
new and used are mixedthat dealer's 144 rows were 62 new against 82 used with no stock-type filter available, so the weighting question is answered wrong unless you separate them first
the gapthe response carries a dealer id but no dealer name, so you know the lot, not the sign
why firsta cent buys a rival's whole stock list, which is the cheapest competitive read in the aisle
$0.04the search that finds the id, then the lot
Everything for sale near you
Say this to your agent
>search carfax used listings within <miles> of <zip> on monid, no make filter, and tell me how many there are in total and how the body types split
What the agent runsdone
carfax /search_used_car_listings$0.01
carfax /search_cars_by_body_type · not run$0.01
carfax /search_new_car_listings · not run$0.01
What came backCharlotte NC · 2026-09-24
in the radius
2015 Jeep Wrangler
Southern Car Connection · 132,713 miles
in the radius
2013 Jeep Wrangler
Tindol Ford Subaru · 113,615 milesthe market10,382 used listings within 25 miles, across 50 pages
per rowdealer name, body type and the CARFAX value badge come back on every listing
the older enda second call filtered to 2017 and earlier surfaced mileage up to 330,131
the readone cent sizes the whole market before you spend anything on a single unit
$0.01per search, 25 of 10,382
Is this unit priced wrong
Say this to your agent
>diff the carfax value response for one <vin> against the search row you already have before you run it on the rest: on one market every key in it was already in that row. if it does add something, put it next to the asking price and read the arrows that explain the gap. if the value comes back null, say so and fall through to comps: on a near-new car it was null and still billed
What the agent runsdone
carfax /get_carfax_value$0.01
kbb /get_car_ratings$0.01
What came backCharlotte NC · 2026-09-24
One unit, three pricesUS dollars
the unit
2008 Chevrolet Corvette
Morrison Motor Cars · 8,695 milesthe gapasking $36,995 against a CARFAX value of $39,530, so $2,535 under
the reasonthe same response gives the why: no accidents, one owner, personal use
the modela second cent buys the book rating, 4.8 out of 5 for that model year
it is not always thereon a second market three of four VINs returned a valuation and the fourth, a near-new car, came back with the value, the badge and the arrows all null and billed the same cent
and check it is not already in the rowon a third market every key it returned was byte-identical to the search row already paid for, so diff it once against a row you own before you run it per VIN; if it matches, the search row is the cheaper source of the same number
the caveatacross 50 sampled listings every badge was GREAT or GOOD, so overpriced units rarely reach the search results
$0.01per VIN
What one actually sold for
Say this to your agent
>search bid.cars on monid for the same model, then pull the similar lots and the sales history for the closest one. there is no location filter on this endpoint, so check where each lot actually is before you call it a comparable, and expect salvage and insurance lots rather than retail cars. the retail vin returns nothing, though it costs nothing either
What the agent runsdone
bid.cars /search_by_keyword$0.01
bid.cars /get_similar_lots$0.01
bid.cars /get_sales_history$0.01
bid.cars /get_vehicle_by_vin · not run$0.02
What came backCharlotte NC · 2026-09-24
the salea comparable unit sold at auction for $30,250, sold by an insurer
the compsrecent sales of the same model at $31,250, $36,750 and $23,900
the spread$6,745 between that auction price and the $36,995 retail ask
no geography at allthis endpoint takes no location parameter, so on a second market none of fifty results were in the state asked for and the top hit was two thousand miles away
and they are not the same kind of carthe lots that came back were flood and front-end damage at six to eight thousand dollars, set against a forty-eight thousand dollar retail asking price
a zero result can still billthe exact retail VIN returned nothing and billed a cent on this page's run, and returned nothing with zero billed units on a later one, so do not treat it as a free VIN check; retail and salvage VINs are separate populations
$0.03three calls
Is this dealer worth dealing with
Say this to your agent
>pull the dealer details from carfax on monid using a VIN from that dealer's lot: the endpoint takes a vin and not a dealer name or id. expect its rating to differ slightly from the other provider's for the same store
What the agent runsdone
carfax /get_dealer_details$0.01
What came backCharlotte NC · 2026-09-24
the score3.9 out of 5 across 21 reviews
the badgea CARFAX Advantage Dealer badge, plus the full opening hours
the sample21 reviews is a small sample; read it as a flag, not a verdict
the two ratings differon a second store this provider said 4.5 and the other said 4.6 for the same dealer, so quote which one you used
the pricea cent, which is cheap enough to run on every dealer in the radius search
$0.01per dealer
What the monthly payment looks like
Say this to your agent
>from the carfax value response for <vin>, give me the estimated monthly payment and the terms it assumes. do not make a second call for it
What the agent runsdone
carfax /get_carfax_value$0.01
What came backCharlotte NC · 2026-09-24
the payment$660.86 a month
the terms10% down, 7.1% APR, 60 months, all stated in the response
the pointthis costs nothing extra: it arrives inside the value call you already paid a cent for
the habitread the whole response before you buy a second one
$0.00already in the value call
The detail call that returns less than the search
Say this to your agent
>pull the cars.com listing detail for one unit on monid using the listing id from the search. then diff it against the search row: on our second market the dealer id, the dealer rating, the review count, the deal rating and the history summary were all null inside the $0.03 detail while the $0.01 search row carried them
What the agent runsdone
cars.com /get_listing_details · not run$0.03
cars.com /search_listings$0.03
What came backCharlotte NC · 2026-09-24
the id is usually thereon a second market the listing id was present on 24 of 24 inventory rows and 24 of 24 search rows, and the $0.03 detail call returned a full record; an earlier version of this job told readers to give up on a call that works
and it carries lessinside that detail response the dealer id, the dealer rating, the review count, the deal rating and the history summary were all null, while the $0.01 search row carried them
when it does fail165 listings across two searches, and the listing id was null on every one
the lot tooall 29 rows of the dealer inventory came back with a null id as well
when it is unreachablethere is no VIN-based lookup, so the only way in is a listing id. On this page's market those came back null, which is why this job never reached the detail; on a second market they were present on all 24 rows and it answered
the cost of finding out$0.06 in searches; the detail attempt itself returned 422 and billed nothing
$0.06two searches, no id to follow
Step 3
Take the cheapest route with Monid.
Every useful call in this aisle is a cent. Flag the mispriced units first, then spend on the few that look wrong.
Radius · $0.01 for the market
Radius search$0.0125 of 10,382
Rival's lot$0.0129 vehicles
Price it · cheapest first
CARFAX value$0.01
Book rating$0.01
Similar lots$0.01
Sales history$0.01
Listing detail$0.03
Check the dealer
Dealer details$0.01
Vehicle details$0.01trim and options · catalog rate
Back to you
rowpriced bycost
1CARFAX value$0.03
2asking price stands$0.01
3Sales history$0.06
4CARFAX value$0.03
5Listing detail$0.09
5 of 5 rows$0.22
Say this to your agent
>tell me which units within <miles> of <zip> are priced away from their carfax value, cheapest source first, and show the cost per unit
“the radius first”one cent returns the whole market with a value badge already on every row
“read the whole response”the monthly payment is inside the value call, so a second call for it is waste
“check where the lot is”the auction endpoint has no location filter at all, so comps come back from anywhere in the country
“diff the detail against the row”the $0.03 detail came back with five dealer fields null that the $0.01 search row carried
One key. 1,700+ tools.
Automotive is the biggest shelf in it.
Listings, values, dealerscarfax · $0.01 / call
A dealer's whole lotcars.com · $0.01 to $0.03
Auction lots and salesbid.cars · $0.01 to $0.02
Book values and ratingskbb · $0.01 / call
Vehicle detailsedmunds · $0.01 / call
Fuel prices by metroaaa · $0.01 to $0.02
The market around the lotzillow · $0.01 to $0.05
Who lives nearbyhomes.com · $0.01 to $0.03
The lot as real estateloopnet · $0.01 to $0.05
Weather that moves stockweather underground · $0.01 to $0.02
Forecast by locationaccuweather · $0.01 to $0.05
Air qualityiqair · $0.01 / call
State dealer statutesvaquill · to $0.06
+ 1,700 more across 55 providers
Browse the catalog →Read more
View all →
Cheapest source first, then fall through
The same routing idea on contact data, with the receipt for every row.
Extract page content for RAG
When a unit only exists on a page, reading the page is the next call.
When the site will not let you in
A browser that signs in and clicks, for the inventory no endpoint covers.