Agent MCP Directory

ScrapingBee: MCP server, CLI and agent surfaces

Yes: ScrapingBee runs a public MCP server at mcp.scrapingbee.com that listed 18 tools to a client with no credentials on 4 November 2026. Examples: get_amazon_search_results, get_amazon_product_details, get_amazon_pricing. It speaks streamable-http, protocol 2025-11-25. Ora's cached score for scrapingbee.com is 61, grade C, read on 5 November 2026.

Fact sheet

CompanyScrapingBee
Domainscrapingbee.com
CategoryDeveloper Tools (assigned by this directory from the registry description)
MCP serverVerified (No authentication)
Ora score61, grade C; Ora scanned the domain on 3 November 2026 and the cached score was read on 5 November 2026 from its score API (Ora's score page for scrapingbee.com, a vendor-published score)
Company found throughthe GitHub MCP Registry; the official MCP Registry
Last verified4 November 2026

MCP server

Endpointhttps://mcp.scrapingbee.com/mcp
How it was foundlisted in the GitHub MCP Registry; the official MCP Registry
Transportstreamable-http
HTTP status200
Protocol version2025-11-25
ServerScrapingBee 2.2.1
Capabilitiesexperimental, logging, prompts, resources, tools, extensions
Tools18
Resources0
Prompts0
Connected4 November 2026, 20:04 UTC
Handshake time1,066 ms over 1 attempt

Tools listed

18 tools

  • get_amazon_search_results
  • get_amazon_product_details
  • get_amazon_pricing
  • ask_chatgpt
  • fast_search
  • ask_gemini
  • get_google_search_results
  • get_page_text
  • get_page_html
  • extract_page_data
  • get_screenshot
  • get_file
  • get_scrapingbee_usage
  • get_walmart_search_results
  • get_walmart_product_details
  • get_youtube_search_results
  • get_youtube_video_metadata
  • get_youtube_video_subtitles

Connector log

2 connections on 4 November 2026, read-only methods only

  • 19:59:21Z https://mcp.scrapingbee.com/mcp sent initialize, notifications/initialized, tools/list, resources/list, prompts/list result verified (HTTP 200) 1301 ms
  • 20:04:23Z https://mcp.scrapingbee.com/mcp sent initialize, notifications/initialized, tools/list, resources/list, prompts/list result verified (HTTP 200) 1066 ms

Authentication

Auth typeNo authentication

The sign-in route, step by step

  1. Send the handshake to https://mcp.scrapingbee.com/mcp over streamable-http. Servers on the 2026-07-28 revision answer server/discover; older servers answer initialize.
  2. Send tools/list, then resources/list and prompts/list if the capabilities name them.
  3. Send no token. The connector sent none and the server answered, so any MCP client can read this list.
  4. Calling a tool is a separate step this directory never takes.

See OAuth for MCP: client ID metadata documents and dynamic registration for how each route works.

CLI

No CLI is on record for ScrapingBee. The directory checks a package only where a source names one, so this is a floor: a company can ship a command line that no source read for this edition mentions. See shipping a CLI agents can use.

Other agent surfaces

Agent surfaces other than MCP and CLI
SurfaceResult
OpenAPI fileNo OpenAPI file at the root paths checked.
WebMCP in the initial HTMLNot checked: the home page was not read by the shared crawl.
MCP discovery fileNot checked: the domain is not in the shared crawl for this edition.
A2A agent cardNot checked: the domain is not in the shared crawl for this edition.
A2A agent card (older path)Not checked: the domain is not in the shared crawl for this edition.
agents.jsonNot checked: the domain is not in the shared crawl for this edition.
OAuth authorization server metadataNot checked: the domain is not in the shared crawl for this edition.
NLWeb endpointNot checked in this edition.

Registry listings

Registry listings
RegistryListed
Official MCP RegistryYes: com.scrapingbee/mcp
GitHub MCP RegistryYes
awesome-remote-mcp-serversNo
SmitheryNot checked: Smithery's listing pages and search API were unavailable on 4 November 2026.
GlamaNot checked: Glama's API needs a key and attribution terms.