Skip to content
Docs

MCP server

Connect Claude, ChatGPT, Cursor or any MCP client to ScraperCompany with OAuth or an API key: hotel, flight and rental data as agent tools, in groups.

The ScraperCompany MCP server exposes the API's hotel, flight and vacation-rental rates, guest reviews, rate monitoring and search data as tools an AI agent can call. Any Model Context Protocol client works: Claude Code, Claude Desktop and claude.ai, ChatGPT, Cursor and others. Sign in through your browser with OAuth, or send your API key as a header.

Overview

Endpoint
https://api.scrapercompany.com/mcp for API users. Assistants such as claude.ai use https://mcp.scrapercompany.com/mcp, which serves the read-only travel set.
Choosing tools
215 tools in 29 groups. A plain connection gets the default set of 24 hotel and flight tools; add groups in the URL, for example https://api.scrapercompany.com/mcp/g/default,reviews.
Transport
Streamable HTTP (MCP 2025-11-25; 2025-06-18 and 2025-03-26 are also accepted). Stateless JSON responses: no sessions and no server-sent stream. The older HTTP+SSE transport (2024-11-05) is not served.
Auth
OAuth 2.1 (browser sign-in, for claude.ai, ChatGPT and any client that can run the flow), or your API key as x-api-key: YOUR_API_KEY or Authorization: Bearer YOUR_API_KEY (details).
Billing
Each tool call is the API call it wraps: priced, rate-limited and recorded in your request history exactly as that call would be. Failed or empty results are free.
Read and write
204 tools only read. 11 tools create, change, run or delete something in your account: the monitoring and webhook groups, which are never in the default set.

Connect a client

Pick the groups you want in the URL (https://api.scrapercompany.com/mcp/g/default,monitoring); plain /mcp is the default set. The same sign-in covers every group URL.

Claude Code

With an API key (the key is read from your environment when you run the command):

Terminal
claude mcp add --transport http scrapercompany https://api.scrapercompany.com/mcp \
  --header "x-api-key: $SCRAPERCOMPANY_API_KEY"

To choose groups, put them in the path. This one adds the guest-review tools to the default set:

Terminal
claude mcp add --transport http scrapercompany https://api.scrapercompany.com/mcp/g/default,reviews \
  --header "x-api-key: $SCRAPERCOMPANY_API_KEY"

Without a key, Claude Code signs you in through the browser: add the server with no header, then finish the sign-in from /mcp inside Claude Code.

Terminal
claude mcp add --transport http scrapercompany https://api.scrapercompany.com/mcp

claude.ai, Claude Desktop and ChatGPT

These connect with OAuth, so there is no key to paste. In Claude, add ScraperCompany from the Claude directory (listed as a Community connector), or add a custom connector with the URL https://mcp.scrapercompany.com/mcp. It lists the read-only travel set, which suits a chat assistant. If that address does not resolve from your network yet, use https://api.scrapercompany.com/mcp/g/travel instead. It is the same server with the same tools, and it works with the same sign-in steps.

For other groups, or in ChatGPT (an MCP server app in developer mode), use https://api.scrapercompany.com/mcp/g/default or the groups you want, and choose OAuth for authentication. Leave any OAuth client fields empty: the client registers itself, or identifies itself with a client metadata URL, then sends you to the sign-in and consent page and connects. The monitoring and webhook tools are marked as writing, so a client that asks before running write tools will ask before running them.

Cursor and other header-capable clients

In ~/.cursor/mcp.json (or your client's equivalent config), point a server at the URL and pass the key as a header. Use the header form in Cursor: the OAuth server only accepts https redirects (and http on localhost), so a client that registers a custom-scheme redirect such as cursor:// is refused at registration.

mcp.json
{
  "mcpServers": {
    "scrapercompany": {
      "url": "https://api.scrapercompany.com/mcp/g/default,monitoring",
      "headers": {
        "x-api-key": "YOUR_API_KEY"
      }
    }
  }
}

Claude Desktop with a key, and stdio-only clients

Clients that only launch local (stdio) servers connect through the mcp-remote bridge. In claude_desktop_config.json:

claude_desktop_config.json
{
  "mcpServers": {
    "scrapercompany": {
      "command": "npx",
      "args": ["-y", "mcp-remote", "https://api.scrapercompany.com/mcp",
               "--header", "x-api-key:${SCRAPERCOMPANY_API_KEY}"],
      "env": { "SCRAPERCOMPANY_API_KEY": "YOUR_API_KEY" }
    }
  }
}

Check the connection

Listing the tools is free and needs no session. Any HTTP client gets the same JSON-RPC reply an MCP client does:

cURL
curl -X POST "https://api.scrapercompany.com/mcp" \
  -H "x-api-key: $SCRAPERCOMPANY_API_KEY" \
  -H "Content-Type: application/json" \
  -H "Accept: application/json, text/event-stream" \
  -d '{"jsonrpc": "2.0", "id": 1, "method": "tools/list"}'

A 401 means the key or token is missing or wrong. Treat the key like a password: keep it in the client's secret store or an environment variable rather than in a config file you share. The MCP Inspector works too: transport Streamable HTTP, the URL above, and either a header or its OAuth settings.

Tool groups

Every tool belongs to one group. A tool list costs the agent context in every conversation, so a connection only sees the groups it asks for.

HowExample
In the path (works for every client, including ones that drop query strings)
https://api.scrapercompany.com/mcp/g/rates,reviews
In the query string
https://api.scrapercompany.com/mcp?groups=rates,reviews
In a header
X-MCP-Groups: rates,reviews

The path form is its own URL, and the query string wins over the header. default is the default set and combines with other groups (default,monitoring); travel is the read-only travel set; all is every tool. An unknown group name is a 400 (JSON-RPC -32602) that names the known ones, not a silently shorter list. A call to a tool outside the connection's groups is a -32602 error that names the group to add, and the initialize instructions tell the agent which groups are on and which exist, so it can ask you to add one.

The read-only travel set

travel is a curated set of 55 tools that only read. It covers hotel search, rates, details and reviews, Airbnb and Vrbo, flights and your credit balance. Nothing in it creates, changes or deletes anything. A connection to https://mcp.scrapercompany.com/mcp that names no groups gets this set. Name it anywhere with https://api.scrapercompany.com/mcp/g/travel. On api.scrapercompany.com, a plain /mcp keeps the default set.

ToolCredits
resolve_hotelFind a hotel
2
find_hotel_everywhereFind a hotel on every booking site
Costs 18 credits by default (Google 9, official site 2, other sources 1 each); free if nothing is found.
hotel_destination_searchSearch hotels in a destination
Costs 4 credits per page; a page with no results is free.
tripcom_find_hotelFind a hotel on Trip.com
2
rakuten_find_hotelFind a hotel on Rakuten Travel
1
jalan_find_hotelFind a hotel on Jalan
1
ikyu_find_hotelFind a hotel on Ikyu
1
yahoo_travel_find_hotelFind a hotel on Yahoo! Travel
1
yanolja_find_stayFind a stay on NOL (Yanolja)
1
lastminute_find_hotelFind a hotel on lastminute.com
1
hotel_rate_calendarHotel nightly rate calendar
8
hotel_offersHotel offers by OTA
Costs 10 credits, or 14 when require_rooms is set.
compare_ota_ratesCompare OTA rates
30
expedia_ratesExpedia rooms and rates
12
hotel_stay_priceGoogle price for one stay
5
hotel_room_matrixRoom types by OTA
5
trivago_offersTrivago offers
12
zenhotels_ratesZenHotels rooms and rates
12
yanolja_ratesNOL (Yanolja) rooms and plans
12
lastminute_pricelastminute.com headline price
12
hostelworld_ratesHostelworld beds and rooms
12
rakuten_ratesRakuten Travel rooms and plans
12
jalan_ratesJalan plans and rooms
12
ikyu_ratesIkyu rooms and plans
12
yahoo_travel_ratesYahoo! Travel rooms and plans
12
official_site_ratesHotel's own-site rates
4
airbnb_searchSearch Airbnb
12
airbnb_priced_calendarAirbnb priced availability calendar
Costs 12 credits plus 1 per night actually priced (at most 92).
vrbo_searchSearch Vrbo
2
vrbo_quoteVrbo stay quote
12
vrbo_calendarVrbo availability calendar
Costs 3 credits (5 for more than 12 months); a listing_id adds 24 the first time it is resolved. Failed or empty answers are free.
hotel_guest_reviewsGoogle Hotels guest reviews
Costs 2 credits per page fetched; failed requests free.
hotel_reviews_all_sourcesGuest reviews across OTAs
Costs 2 credits per page per source; failed pages free.
review_insightsReview response insights
Costs 2 credits per page per source; failed pages free.
booking_reviewsBooking.com reviews
2
expedia_reviewsExpedia reviews
2
hotelscom_reviewsHotels.com reviews
2
tripadvisor_reviewsTripadvisor reviews
1
agoda_reviewsAgoda reviews
2
airbnb_reviewsAirbnb reviews
2
vrbo_reviewsVrbo reviews
2
tripcom_reviewsTrip.com reviews
2
hotel_detailsGoogle Hotels property details
5
flight_searchSearch flights
4
flight_next_legFlight return or next leg
4
flight_booking_optionsFlight booking options
5
flight_location_searchFind airport or city codes
2
flight_price_calendarFlight price calendar
6
credit_balanceCredit balance
free
discovercars_locationsFind a DiscoverCars pick-up place
1
discovercars_offersDiscoverCars car-rental offers
10
sixt_locationsFind a Sixt branch
1
sixt_offersSixt car-rental offers
4
attractions_calendarTheme-park ticket calendar
Costs 4 credits per started 31 days of range, plus 2 per 31 days with include_capacity (not charged when no date has capacity). Free when no date is priced.
attraction_parksTheme parks covered
free
GroupToolsIn the default setWhat
default
24
all
The hotel essentials (search, rates, own-site rates, rentals, the cross-source review tools, credits and the catalog) and the flight tools.
2
2
Credit balance and the endpoint catalog
19
6
Hotel name or link to each site's id; destination search
21
4
Google and OTA (Booking, Hotels.com, Expedia, Agoda, Tripadvisor, Trivago, ZenHotels, Hostelworld, Rakuten Travel, Jalan, Ikyu, Yahoo! Travel, Yanolja, lastminute.com) prices for a stay or a horizon
3
1
The hotel's own booking engine: rates and loyalty ladder
5
4
Airbnb and Vrbo search, quotes and calendars
11
2
Guest reviews per source, merged, and response insights
10
none
Property details, photos, facilities, nearby places
18
none
Compsets: scheduled rate shopping, parity, history, alerts, exports, the free trial monitor (changes data)
4
none
Webhook endpoints for rate and quota events (changes data)
7
5
Google Flights search, legs, booking options, calendars
7
none
Car-rental offers (DiscoverCars, Sixt) and their place and branch lookups, plus AutoTrader and AutoScout24 vehicle listings
2
none
Theme-park and attraction day-ticket prices per visit date, and the parks covered
8
none
Google Search, News, Maps, Trends and Autocomplete results
8
none
LinkedIn, Indeed, Upwork, Google Jobs and company careers-page openings
3
none
Facebook public events: keyword search, listing pages, event detail
10
none
YouTube and TikTok: search, video, transcript, comments, channel, profile and hashtag
6
none
Google Play and Apple App Store: app details, reviews, search
20
none
Amazon, Walmart, Google Shopping, TikTok Shop US and Shopify: products, sellers, reviews and merchant discovery
12
none
Property listings for sale or rent: Domain.com.au, Redfin, Zillow, Centris (Quebec) and StreetEasy (New York City), with location lookup
9
none
Company research: SEC filings, financials, insider trades, 13F holdings, private raises (Form D), the earnings calendar, businesses for sale
4
none
Second-hand and handmade listings: Craigslist, Kijiji, Karrot and Etsy
4
none
FAA aircraft registration, recent documents and registry search by owner, maker, model or place
3
none
Yelp local-business search, details and recent reviews
3
none
Ticketing and live events: SeatGeek (with live price stats), Ticketmaster and StubHub
2
none
Event listings by city or area: a unified events_search across Eventbrite, Luma, Resident Advisor, EventsEye trade fairs, Songkick, RunSignup, sports schedules, Eventim, Bandsintown, Live Nation, Dice, AllEvents and university calendars, plus Eventbrite event detail
1
none
WhatsApp number registration and public-profile checks
5
none
Live vessel positions near a point (MarineTraffic), and cruise ships: cruise search with itineraries, ship details, live cruise-ship positions and port schedules (CruiseMapper)
7
none
Wholesale sourcing and trade records: 1688.com products and suppliers, US import bills of lading by company, supplier or HS code (ImportYeti), ThomasNet industrial suppliers by product, service or company name
1
none
Any public web page as markdown, text or HTML (generic scrape)
all
215
all
Everything.

A group marked “none” in the third column is opt-in: a plain connection never lists its tools. Name it to get it, alone or next to the default set, for example https://api.scrapercompany.com/mcp/g/default,serp,jobs for the hotel and flight tools plus Google results and jobs.

Tools

Each tool wraps one REST endpoint and takes that endpoint's request fields (most of them), validated the same way. Options that change a call's price, such as basis and verify_sources on the calendar, stay REST-only. Tool descriptions carry the cost too, so the agent sees what a call will charge before making it. api_catalog is the discovery tool: an agent that needs a capability no listed tool names can read the full REST inventory, the cost of each endpoint and the shared conventions from it.

A typical hotel conversation: resolve_hotel with a name and city returns a token; hotel_rate_calendar with that token and days: 60 returns 60 nightly prices; then hotel_offers for the cheapest night returns each booking site's price. For flights: flight_search with a return date, flight_next_leg with the chosen outbound's departure_token, then flight_booking_options with the return's booking_token.

Endpoints without a reference page

52 tools below wrap an endpoint that is not in the public API reference yet (the monitoring and webhook tools and some review and content tools). They are shown without a link; the tool's own input schema, from tools/list, is the specification.

Read-only and write tools

Every tool states a title and its readOnlyHint, destructiveHint and openWorldHint. Search, rate, review, content and catalog tools only read (readOnlyHint: true, destructiveHint: false). The monitoring and webhook tools that create, change, run or delete something carry readOnlyHint: false, and the deletes and compset_update (which overwrites settings) carry destructiveHint: true. Tools that only touch your own stored objects carry openWorldHint: false. compset_export_link returns a download link instead of the file: a URL under /mcp/exports/ that needs no key, serves that one export as the key that made it, and expires after 15 minutes or when that key is revoked.

catalog (2)

Credit balance and the endpoint catalog.

ToolWrapsCreditsKind
credit_balance(default set)Credit balance
free
read
api_catalog(default set)What this API can do
GET /v1
free
read

Hotel name or link to each site's id; destination search.

ToolWrapsCreditsKind
resolve_hotel(default set)Find a hotel
2
read
resolve_google_hotel(default set)Find a hotel by address
12
read
find_hotel_everywhere(default set)Find a hotel on every booking site
POST /v1/meta/search
Costs 18 credits by default (Google 9, official site 2, other sources 1 each); free if nothing is found.
read
hotel_link_ids(default set)Identify a hotel link
POST /v1/meta/extract-id
Costs 1 credit only to follow a short link; else free.
read
hotel_destination_search(default set)Search hotels in a destination
Costs 4 credits per page; a page with no results is free.
read
expedia_find_hotel(default set)Find a hotel on Expedia
2
read
booking_find_hotelFind a hotel on Booking.com
4
read
booking_hotel_searchSearch a destination on Booking.com
Costs 4 credits per started 25 properties returned.
read
hotelscom_find_hotelFind a hotel on Hotels.com
2
read
agoda_find_hotelFind a hotel on Agoda
2
read
tripadvisor_find_hotelFind a place on Tripadvisor
2
read
tripcom_find_hotelFind a hotel on Trip.com
POST /v1/ota/tripcom/search
2
read
rakuten_find_hotelFind a hotel on Rakuten Travel
1
read
jalan_find_hotelFind a hotel on Jalan
1
read
ikyu_find_hotelFind a hotel on Ikyu
1
read
yahoo_travel_find_hotelFind a hotel on Yahoo! Travel
1
read
yanolja_find_stayFind a stay on NOL (Yanolja)
1
read
lastminute_find_hotelFind a hotel on lastminute.com
1
read
booking_destination_lookupBooking.com destination lookup
POST /v1/ota/booking/destinations
2
read

rates (21)

Google and OTA (Booking, Hotels.com, Expedia, Agoda, Tripadvisor, Trivago, ZenHotels, Hostelworld, Rakuten Travel, Jalan, Ikyu, Yahoo! Travel, Yanolja, lastminute.com) prices for a stay or a horizon.

ToolWrapsCreditsKind
hotel_rate_calendar(default set)Hotel nightly rate calendar
8
read
hotel_offers(default set)Hotel offers by OTA
Costs 10 credits, or 14 when require_rooms is set.
read
compare_ota_rates(default set)Compare OTA rates
30
read
expedia_rates(default set)Expedia rooms and rates
12
read
hotel_stay_priceGoogle price for one stay
5
read
hotel_room_matrixRoom types by OTA
5
read
booking_rate_calendarBooking.com price calendar
12
read
booking_roomsBooking.com rooms and rates
12
read
hotelscom_priceHotels.com price for a stay
12
read
hotelscom_roomsHotels.com rooms and rates
12
read
agoda_ratesAgoda rooms and rates
12
read
tripadvisor_offersTripadvisor offers
12
read
trivago_offersTrivago offers
12
read
zenhotels_ratesZenHotels rooms and rates
12
read
yanolja_ratesNOL (Yanolja) rooms and plans
12
read
lastminute_pricelastminute.com headline price
12
read
hostelworld_ratesHostelworld beds and rooms
12
read
rakuten_ratesRakuten Travel rooms and plans
12
read
jalan_ratesJalan plans and rooms
12
read
ikyu_ratesIkyu rooms and plans
12
read
yahoo_travel_ratesYahoo! Travel rooms and plans
12
read

official (3)

The hotel's own booking engine: rates and loyalty ladder.

ToolWrapsCreditsKind
official_site_rates(default set)Hotel's own-site rates
4
read
official_loyalty_ladderBrand.com cash vs member vs points
POST /v1/official/loyalty
4
read
official_enginesNumber of supported booking engines
GET /v1/official/engines
free
read

rentals (5)

Airbnb and Vrbo search, quotes and calendars.

ToolWrapsCreditsKind
airbnb_search(default set)Search Airbnb
12
read
airbnb_priced_calendar(default set)Airbnb priced availability calendar
Costs 12 credits plus 1 per night actually priced (at most 92).
read
vrbo_search(default set)Search Vrbo
2
read
vrbo_quote(default set)Vrbo stay quote
12
read
vrbo_calendarVrbo availability calendar
Costs 3 credits (5 for more than 12 months); a listing_id adds 24 the first time it is resolved. Failed or empty answers are free.
read

reviews (11)

Guest reviews per source, merged, and response insights.

ToolWrapsCreditsKind
hotel_guest_reviews(default set)Google Hotels guest reviews
Costs 2 credits per page fetched; failed requests free.
read
hotel_reviews_all_sources(default set)Guest reviews across OTAs
POST /v1/reviews
Costs 2 credits per page per source; failed pages free.
read
review_insightsReview response insights
POST /v1/reviews/insights
Costs 2 credits per page per source; failed pages free.
read
booking_reviewsBooking.com reviews
2
read
expedia_reviewsExpedia reviews
2
read
hotelscom_reviewsHotels.com reviews
2
read
tripadvisor_reviewsTripadvisor reviews
1
read
agoda_reviewsAgoda reviews
2
read
airbnb_reviewsAirbnb reviews
2
read
vrbo_reviewsVrbo reviews
2
read
tripcom_reviewsTrip.com reviews
2
read

content (10)

Property details, photos, facilities, nearby places.

ToolWrapsCreditsKind
hotel_detailsGoogle Hotels property details
5
read
hotel_photosGoogle Hotels photos
POST /v1/hotels/photos
3
read
hotel_nearbyPlaces near a hotel
POST /v1/hotels/nearby
3
read
booking_detailsBooking.com property details
POST /v1/ota/booking/details
3
read
booking_photosBooking.com photos
POST /v1/ota/booking/photos
3
read
booking_facilitiesBooking.com facilities
POST /v1/ota/booking/facilities
3
read
booking_review_summaryBooking.com review summary
POST /v1/ota/booking/review-summary
3
read
tripadvisor_detailsTripadvisor place details
2
read
airbnb_listingAirbnb listing details
POST /v1/ota/airbnb/listing
3
read
airbnb_photosAirbnb listing photos
POST /v1/ota/airbnb/photos
3
read

monitoring (18)

Compsets: scheduled rate shopping, parity, history, alerts, exports, the free trial monitor (changes data).

ToolWrapsCreditsKind
compset_estimatePrice a compset before creating it
POST /v1/compsets/estimate
free
read
compset_createCreate a compset
POST /v1/compsets
Free; its runs bill the calls that price.
write
compset_listList compsets
GET /v1/compsets
free
read
compset_getRead a compset
GET /v1/compsets/{compset_id}
free
read
compset_updateChange a compset
PATCH /v1/compsets/{compset_id}
free
destructive
compset_add_membersAdd hotels to a compset
POST /v1/compsets/{compset_id}/members
free
write
compset_remove_memberRemove a hotel from a compset
DELETE /v1/compsets/{compset_id}/members/{member_id}
free
destructive
compset_deleteDelete a compset
DELETE /v1/compsets/{compset_id}
free
destructive
compset_runShop a compset now
POST /v1/compsets/{compset_id}/run
Free to call; each upstream call that prices bills at its REST price (compset_estimate says how much). Sold-out, empty and failed calls are free.
write
compset_run_statusRead a compset run
GET /v1/compsets/{compset_id}/runs/{run_id}
free
read
compset_ratesCompset rate grid
GET /v1/compsets/{compset_id}/rates
free
read
compset_parityCompset parity and positioning
GET /v1/compsets/{compset_id}/parity
free
read
compset_historyCompset rate history
GET /v1/compsets/{compset_id}/history
free
read
compset_export_linkExport link for a compset
GET /v1/compsets/{compset_id}/export
free
read
compset_alertsList compset alert rules
GET /v1/compsets/{compset_id}/alerts
free
read
compset_alert_createAdd a compset alert rule
POST /v1/compsets/{compset_id}/alerts
free
write
compset_alert_deleteDelete a compset alert rule
DELETE /v1/compsets/{compset_id}/alerts/{rule_id}
free
destructive
monitor_trialStart the free 48-hour monitor
POST /v1/monitor/trial
free
write

webhooks (4)

Webhook endpoints for rate and quota events (changes data).

ToolWrapsCreditsKind
webhook_event_typesWebhook event types
GET /v1/webhooks/events
free
read
webhook_listList webhooks
GET /v1/webhooks
free
read
webhook_createCreate a webhook
POST /v1/webhooks
free
write
webhook_testSend a test event
POST /v1/webhooks/{endpoint_id}/test
free
write

flights (7)

Google Flights search, legs, booking options, calendars.

ToolWrapsCreditsKind
flight_search(default set)Search flights
4
read
flight_next_leg(default set)Flight return or next leg
4
read
flight_booking_options(default set)Flight booking options
5
read
flight_location_search(default set)Find airport or city codes
2
read
skyscanner_searchSkyscanner flight search
POST /v1/skyscanner/search
45
read
skyscanner_pricesSkyscanner indicative fares
POST /v1/skyscanner/prices
2
read
flight_price_calendar(default set)Flight price calendar
6
read

cars (7)

Car-rental offers (DiscoverCars, Sixt) and their place and branch lookups, plus AutoTrader and AutoScout24 vehicle listings.

ToolWrapsCreditsKind
discovercars_locationsFind a DiscoverCars pick-up place
1
read
discovercars_offersDiscoverCars car-rental offers
10
read
autoscout24_detailAutoScout24 listing detail
2
read
sixt_locationsFind a Sixt branch
1
read
sixt_offersSixt car-rental offers
4
read
autotrader_searchAutoTrader vehicle search (US)
2
read
autoscout24_searchAutoScout24 vehicle search (Europe)
2
read

attractions (2)

Theme-park and attraction day-ticket prices per visit date, and the parks covered.

ToolWrapsCreditsKind
attractions_calendarTheme-park ticket calendar
Costs 4 credits per started 31 days of range, plus 2 per 31 days with include_capacity (not charged when no date has capacity). Free when no date is priced.
read
attraction_parksTheme parks covered
free
read

serp (8)

Google Search, News, Maps, Trends and Autocomplete results.

ToolWrapsCreditsKind
google_searchGoogle web search
4
read
google_newsGoogle News stories
2
read
maps_searchGoogle Maps places
3
read
maps_leadsGoogle Maps business leads
6
read
maps_place_detailGoogle Maps place detail
3
read
trends_interestGoogle Trends interest over time
2
read
trends_relatedGoogle Trends related queries
2
read
google_autocompleteGoogle autocomplete suggestions
2
read

jobs (8)

LinkedIn, Indeed, Upwork, Google Jobs and company careers-page openings.

ToolWrapsCreditsKind
linkedin_jobsSearch LinkedIn jobs
2
read
upwork_jobsSearch Upwork jobs
2
read
upwork_jobUpwork job details
2
read
linkedin_jobLinkedIn job posting
2
read
indeed_jobsSearch Indeed jobs
2
read
careers_jobsA company's open jobs
2
read
careers_detectDetect a careers page's job system
1
read
google_jobsGoogle Jobs listings
3
read

facebook (3)

Facebook public events: keyword search, listing pages, event detail.

ToolWrapsCreditsKind
facebook_events_searchSearch Facebook events
POST /v1/facebook/events/search
2
read
facebook_eventsFacebook events listing page
2
read
facebook_eventFacebook event details
2
read

video (10)

YouTube and TikTok: search, video, transcript, comments, channel, profile and hashtag.

ToolWrapsCreditsKind
youtube_searchSearch YouTube
2
read
youtube_videoYouTube video details
2
read
youtube_transcriptYouTube transcript
2
read
youtube_commentsYouTube comments
2
read
youtube_channelYouTube channel
2
read
tiktok_videoTikTok video
2
read
tiktok_profileTikTok profile
Costs 3 credits; 2 when the page embeds the account
read
tiktok_searchTikTok search
3
read
tiktok_hashtagTikTok hashtag
3
read
tiktok_commentsTikTok comments
3
read

apps (6)

Google Play and Apple App Store: app details, reviews, search.

ToolWrapsCreditsKind
google_play_appGoogle Play app listing
2
read
google_play_reviewsGoogle Play reviews
2
read
google_play_searchSearch Google Play
2
read
app_store_appApp Store app record
2
read
app_store_reviewsApp Store reviews
2
read
app_store_searchSearch the App Store
2
read

shopping (20)

Amazon, Walmart, Google Shopping, TikTok Shop US and Shopify: products, sellers, reviews and merchant discovery.

ToolWrapsCreditsKind
amazon_searchSearch Amazon
2
read
amazon_reviewsAmazon ratings and reviews
2
read
amazon_productAmazon product details
POST /v1/amazon/product
2
read
walmart_searchSearch Walmart
2
read
walmart_reviewsWalmart reviews
2
read
target_storesTarget stores near a ZIP
1
read
target_searchSearch Target
6
read
target_productTarget product details
3
read
google_shoppingGoogle Shopping products
3
read
tiktok_shop_productTikTok Shop US product
POST /v1/tiktok/shop/product
2
read
tiktok_shop_sellerTikTok Shop US seller
POST /v1/tiktok/shop/seller
2
read
tiktok_shop_categoryTikTok Shop US category
POST /v1/tiktok/shop/category
2
read
shopify_productsShopify products
Costs 9 credits for a page larger than 50 products, and 3 when limit is 50 or less. Empty pages are free.
read
shopify_productOne Shopify product
2
read
shopify_reviewsShopify product reviews
Costs 9 credits for the product page and 2 per further page of 20. Default max_reviews 50 holds 13. Empty is free.
read
shopify_storeShopify store identity
3
read
shopify_profileShopify store profile
14
read
shopify_searchSearch a Shopify catalog
Costs 2 credits. Empty is free.
read
shopify_stores_searchFind Shopify merchants
POST /v1/retail/shopify/stores/search
Costs 5 credits per full store or 1 per light store, and at least 5 when any store is returned. Empty is free.
read
shopify_stores_lookupLook up Shopify merchants
POST /v1/retail/shopify/stores/lookup
Costs 5 credits per store found. None found is free.
read

property (12)

Property listings for sale or rent: Domain.com.au, Redfin, Zillow, Centris (Quebec) and StreetEasy (New York City), with location lookup.

ToolWrapsCreditsKind
domain_locationsFind a Domain location
1
read
streeteasy_searchStreetEasy NYC listings
Costs 2 credits per page fetched (about 14 rentals or 38 sales a page). No listings is free.
read
streeteasy_listingStreetEasy listing detail
2
read
streeteasy_buildingStreetEasy building detail
2
read
streeteasy_agentStreetEasy agent profile
2
read
redfin_searchRedfin for-sale listings
2
read
zillow_searchZillow for-sale listings
2
read
centris_searchCentris for-sale listings
2
read
redfin_propertyRedfin property detail
2
read
zillow_propertyZillow property detail
2
read
centris_propertyCentris property detail
2
read
domain_listingsDomain.com.au listings
3
read

companies (9)

Company research: SEC filings, financials, insider trades, 13F holdings, private raises (Form D), the earnings calendar, businesses for sale.

ToolWrapsCreditsKind
sec_find_companyFind a company's SEC CIK
1
read
sec_filingsA company's SEC filings
1
read
sec_financialsA company's reported financials
2
read
sec_insider_tradesInsider trades (Form 4)
4
read
sec_fund_holdingsA fund's 13F holdings
4
read
sec_searchSearch SEC filings text
2
read
sec_private_raisesPrivate raises (Form D)
4
read
earnings_calendarEarnings calendar for a day
2
read
businesses_for_saleOnline businesses for sale
2
read

marketplace (4)

Second-hand and handmade listings: Craigslist, Kijiji, Karrot and Etsy.

ToolWrapsCreditsKind
craigslist_searchCraigslist search
2
read
kijiji_searchKijiji search
2
read
karrot_searchKarrot search
2
read
etsy_searchEtsy search
17
read

aviation (4)

FAA aircraft registration, recent documents and registry search by owner, maker, model or place.

ToolWrapsCreditsKind
aircraft_registrationFAA aircraft registration
2
read
aircraft_documentsFAA recent aircraft documents
2
read
aircraft_recordsFAA aircraft records
4
read
aircraft_searchSearch the FAA aircraft registry
2
read

local (3)

Yelp local-business search, details and recent reviews.

ToolWrapsCreditsKind
yelp_searchYelp business search
POST /v1/yelp/search
5
read
yelp_businessYelp business details
POST /v1/yelp/business
5
read
yelp_reviewsYelp business reviews
POST /v1/yelp/reviews
5
read

events (3)

Ticketing and live events: SeatGeek (with live price stats), Ticketmaster and StubHub.

ToolWrapsCreditsKind
seatgeek_searchSeatGeek event search
2
read
ticketmaster_searchTicketmaster event search
2
read
stubhub_searchStubHub event search
5
read

event_discovery (2)

Event listings by city or area: a unified events_search across Eventbrite, Luma, Resident Advisor, EventsEye trade fairs, Songkick, RunSignup, sports schedules, Eventim, Bandsintown, Live Nation, Dice, AllEvents and university calendars, plus Eventbrite event detail.

ToolWrapsCreditsKind
events_searchEvents by city and source
Costs 2 credits, or 1 credit for sports and livenation. Empty or failed results are free.
read
eventbrite_eventEventbrite event detail
2
read

messaging (1)

WhatsApp number registration and public-profile checks.

ToolWrapsCreditsKind
whatsapp_checkWhatsApp number check
POST /v1/whatsapp/check
2
read

maritime (5)

Live vessel positions near a point (MarineTraffic), and cruise ships: cruise search with itineraries, ship details, live cruise-ship positions and port schedules (CruiseMapper).

ToolWrapsCreditsKind
maritime_vesselsLive vessels near a point
2
read
cruise_searchCruise search (CruiseMapper)
2
read
cruise_shipCruise ship details
3
read
cruise_ship_positionsLive cruise-ship positions
2
read
cruise_port_scheduleCruise port schedule
2
read

trade (7)

Wholesale sourcing and trade records: 1688.com products and suppliers, US import bills of lading by company, supplier or HS code (ImportYeti), ThomasNet industrial suppliers by product, service or company name.

ToolWrapsCreditsKind
alibaba1688_search1688.com wholesale search
2
read
alibaba1688_product1688.com product
10
read
import_records_searchUS import records search
2
read
import_records_profileUS import profile
5
read
thomasnet_searchThomasNet supplier search
22
read
thomasnet_profileThomasNet supplier profile
22
read
import_records_hs_codeUS imports by HS code
3
read

web (1)

Any public web page as markdown, text or HTML (generic scrape).

ToolWrapsCreditsKind
scrape_urlRead a web page
Costs 1 credit; 5 with render, 10 with premium, 75 with both. Non-2xx, blocked, failed and empty pages are free.
read

Credits

A tool call costs exactly what the route it wraps costs (the Credits column above, and the end of every tool description), is free when it fails or comes back empty, and is charged to the same account and ledger whether the connection signed in with the API key or with OAuth. It is a per-route price, not a flat price per call: a name lookup is a couple of credits, a 330-night calendar is 8, and an OTA comparison is 30, so an agent planning a long session should read the cost in each tool description or call credit_balance. initialize, tools/list and ping are free. See Credits for the full schedule.

credit_balance needs a customer account

Like GET /v1/billing, it answers only for keys linked to a customer account; other keys get a tool error with status 404.

OAuth sign-in

/mcp is an OAuth 2.1 protected resource, as the MCP authorization spec describes. A client that calls it without credentials gets a challenge that points at everything it needs:

Challenge
HTTP/1.1 401
WWW-Authenticate: Bearer realm="ScraperCompany",
  resource_metadata="https://api.scrapercompany.com/.well-known/oauth-protected-resource/mcp"
EndpointSpecNotes
GET /.well-known/oauth-protected-resource
RFC 9728
resource and authorization_servers.
GET /.well-known/oauth-authorization-server
RFC 8414
The issuer and the endpoints below.
POST /oauth/register
RFC 7591
Open dynamic registration, for public clients only: token_endpoint_auth_method: none, and no client secret is ever issued.
client_id
Client ID Metadata Documents
A client may skip registration and use an https URL as its client_id. The server fetches the JSON document at that URL, which must name the same URL and list the redirect URIs.
GET, POST /oauth/authorize
OAuth 2.1
Authorization code with PKCE S256, which is required.
POST /oauth/token
OAuth 2.1
authorization_code and refresh_token grants.
POST /oauth/revoke
RFC 7009
Revoking a refresh token ends the whole grant.

All of these are on api.scrapercompany.com, the issuer. A connection to https://mcp.scrapercompany.com/mcp gets a challenge naming that address as the resource and api.scrapercompany.com as the authorization server.

The flow, as you see it:

  1. The client opens /oauth/authorize in your browser.
  2. If you are signed out, choose Continue with Google (which also creates an account if you do not have one), create an account with your email, or sign in with your email and password (the same account as the dashboard). A new email account is asked to confirm its address first; the link in that email brings you straight back to this flow.
  3. The consent page names the application as it declared itself (marked self-declared), the host it will send you back to and your account. If the account has no API key yet, the page says that allowing creates one.
  4. Allow sends you back to the client with a code, which it swaps, with its PKCE verifier, for tokens.

Dashboard > Connected apps lists every application you connected. Disconnect ends that application's access at once.

What a token is, and is not:

  • An access token (mcp_at_…, valid 1 hour) authenticates /mcp only. Sent to a /v1/* route, or as x-api-key, it is an invalid key (401).
  • It acts as your account's current API key: tool calls are rate limited, metered, billed and recorded exactly as with that key, and see the same compsets and webhooks. Rotating the key keeps the connection working. Revoking the key or deactivating the account stops it (401 invalid_token). Only customer accounts can consent; operator accounts cannot.
  • A refresh token (mcp_rt_…, valid 30 days) is single use: each refresh returns a new pair, and presenting a used refresh token again is refused.
  • The scope is mcp (call the tools), and the audience is the /mcp resource you connected to. The group URLs are the same resource, so one sign-in covers them all. A token issued for https://mcp.scrapercompany.com/mcp works only there, and one for https://api.scrapercompany.com/mcp works only on the API host.
  • Redirect URIs must be https, or http on localhost, 127.0.0.1 or [::1]. A loopback redirect may use any port and any of those three names, so command-line clients can pick a free port each run. PKCE S256 is required, and an authorization code lives 5 minutes and works once.

Results

A successful call returns the route's JSON twice: in structuredContent, and as text in content[0].text for clients that only read text. Per-row provenance blocks are left out to save the agent's context; call the REST endpoint when you need them for an audit.

_meta on every result gives the HTTP status of the wrapped call and, when the call was metered, the credits charged and remaining (as strings, copied from the credit headers):

tools/call result (shortened)
{
  "jsonrpc": "2.0",
  "id": 7,
  "result": {
    "content": [{ "type": "text", "text": "{\"token\":\"ChUIoben2Mv6-CYaCi9tLzA3czVwbjQQAQ\",\"requested_days\":60,\"coverage\":0.95,\"rates\":[...]}" }],
    "structuredContent": { "token": "ChUIoben2Mv6-CYaCi9tLzA3czVwbjQQAQ", "requested_days": 60, "coverage": 0.95, "rates": ["..."] },
    "isError": false,
    "_meta": { "http_status": 200, "credits_charged": "8", "credits_remaining": "992" }
  }
}

Errors

Errors come back at two levels.

Tool errors are results with isError: true, so the model can read them and fix its next call. The text is {"error": ..., "status": ...} with the wrapped route's detail and status. They cover:

  • a missing or unknown argument (the message lists the allowed ones);
  • anything the route rejects, such as a date in the past (422) or a property Google no longer lists (404);
  • a compset or webhook that does not exist or is not yours (404);
  • a source that failed or blocked the request (502) or changed its contract (503);
  • no credits left (402, once enforcement is on), which says nothing was charged and gives a one-time checkout link for the smallest credit pack plus the pricing page (connected through ChatGPT, the pricing page only), or the key's rate limit (429).

An unexpected server error is also a tool error, with a generic message. Tool errors are not charged.

Tool error
{
  "jsonrpc": "2.0",
  "id": 8,
  "result": {
    "content": [{ "type": "text", "text": "{\"error\": \"check_in is in the past\", \"status\": 422}" }],
    "isError": true,
    "_meta": { "http_status": 422 }
  }
}

Protocol errors are JSON-RPC errors, for problems with the message itself:

CodeWhen
-32700
The body is not valid JSON (HTTP 400).
-32600
Not a valid JSON-RPC 2.0 request; an unsupported MCP-Protocol-Version (HTTP 400); a browser Origin that is not allowed (HTTP 403); a body over 256 KB (HTTP 413).
-32601
Unknown method. The server implements initialize, ping, tools/list and tools/call.
-32602
Unknown tool, a tool outside the connection's groups, an unknown group name, or params / arguments that are not objects.
-32000
Rate limit exceeded (HTTP 429 with Retry-After: 60 for a single message).

A missing or invalid key or token is an HTTP 401 before any JSON-RPC handling.

Limits

  • One JSON-RPC message per POST under 2025-11-25 and 2025-06-18 (batching was removed in 2025-06-18). 2025-03-26 clients may batch up to 10 messages.
  • Request bodies over 256 KB get 413.
  • Tool calls count against your key's rate limit through the route they wrap. Every other message (initialize, tools/list, ping), and any tool call rejected before it reaches a route, counts once. An OAuth connection shares its account key's limit.
  • A browser Origin must be scrapercompany.com (or a subdomain) or localhost; anything else gets 403, which guards against DNS rebinding. Agents and CLIs send no Origin.
  • GET and DELETE on /mcp (and the group URLs) return 405: the server never opens a stream or a session.

Questions about this page?

Send the page link and your question, and the team will answer.

Email support