generated: '2026-08-29'
method: probed
source: >-
  https://api.scrapingant.com/mcp/ (anonymous JSON-RPC tools/list, HTTP 200) and
  https://docs.scrapingant.com/mcp-server
name: ScrapingAnt MCP Server
provider: ScrapingAnt
providerId: scrapingant
description: >-
  ScrapingAnt operates a first-party hosted MCP server that gives an AI agent live web access
  through the same headless-Chrome cluster and rotating proxy pool that backs the /v2/general
  REST endpoint. Three tools return the same fetched page in three formats - raw HTML, cleaned
  Markdown, and plain text - so the agent can pick the representation that fits the task.
  Authentication is the same x-api-key the REST API uses, sent as an HTTP header, and MCP calls
  draw down the same API-credit pool.
status: live
deployment:
  mode: remote
  endpoint: https://api.scrapingant.com/mcp/
  install: null
  package: null
  auth: api-key
  verified: probed
transport:
  protocol: streamable-http
  declared_as: streamableHttp
  note: >-
    https://api.scrapingant.com/mcp (no trailing slash) answers HTTP 307 and redirects to
    /mcp/. The provider's own published configuration snippets use the un-slashed form for
    Claude Desktop, Cursor, Claude Code and Cline, and the slashed form for VS Code; both
    resolve. tools/list was probed successfully against the slashed URL.
authentication:
  scheme: api-key
  header: x-api-key
  obtained_from: https://app.scrapingant.com/signup
  anonymous_introspection: true
  note: >-
    tools/list returns the complete tool set with full inputSchema WITHOUT credentials - the
    schema surface is open even though tool invocation requires a key. This is the best case
    for agent discovery: a client can plan against the real schemas before a key exists.
clients_documented:
- Claude Desktop
- Claude Code (CLI)
- Cursor
- Windsurf
- Cline
- VS Code / GitHub Copilot
install_examples:
- client: Claude Code
  command: >-
    claude mcp add scrapingant --transport http https://api.scrapingant.com/mcp
    -H "x-api-key: <YOUR-API-KEY>"
  source: https://docs.scrapingant.com/mcp-server
- client: Cursor / Windsurf / Claude Desktop / Cline
  config: |
    {
      "mcpServers": {
        "scrapingant": {
          "url": "https://api.scrapingant.com/mcp",
          "transport": "streamableHttp",
          "headers": { "x-api-key": "<YOUR-API-KEY>" }
        }
      }
    }
  source: https://docs.scrapingant.com/mcp-server
tools:
- name: get_web_page_html
  description: Fetch (scrape) a URL using ScrapingAnt and return the web page content as HTML.
  parameters:
  - name: url
    type: string
    required: true
    description: The URL of the page to extract (scrape).
  - name: browser
    type: boolean
    required: false
    default: true
    description: Whether to use browser rendering.
  - name: proxy_type
    type: string
    required: false
    default: datacenter
    description: >-
      Type of proxy to use. Use 'residential' when facing anti-bot detection.
  - name: proxy_country
    type: string
    required: false
    default: null
    description: >-
      ISO-3166 country code. Published set: ae, au, br, ca, cn, cz, de, es, fr, gb, hk, id,
      il, in, it, jp, kr, my, nl, ph, pl, ru, sa, sg, th, us, vn.
  output: 'string (wrapped by FastMCP as {"result": <string>})'
- name: get_web_page_markdown
  description: Fetch (scrape) a URL using ScrapingAnt and return the web page content as Markdown.
  parameters:
  - name: url
    type: string
    required: true
  - name: browser
    type: boolean
    required: false
    default: true
  - name: proxy_type
    type: string
    required: false
    default: datacenter
  - name: proxy_country
    type: string
    required: false
    default: null
  output: 'string (wrapped by FastMCP as {"result": <string>})'
- name: get_web_page_text
  description: Fetch (scrape) a URL using ScrapingAnt and return the web page content as plain text.
  parameters:
  - name: url
    type: string
    required: true
  - name: browser
    type: boolean
    required: false
    default: true
  - name: proxy_type
    type: string
    required: false
    default: datacenter
  - name: proxy_country
    type: string
    required: false
    default: null
  output: 'string (wrapped by FastMCP as {"result": <string>})'
capabilities:
  tools: true
  resources: false
  prompts: false
  note: >-
    Only tools/list was probed. resources/list and prompts/list were not exercised, so the
    false values above are "not observed", not "confirmed absent".
raw_manifest: mcp/scrapingant-mcp-tools-list.json
implementation:
  framework: FastMCP
  evidence: >-
    Every tool carries an outputSchema titled _WrappedResult with the
    x-fastmcp-wrap-result extension and a _meta._fastmcp block.
billing:
  model: api-credits
  note: >-
    MCP requests consume the same API credits as direct REST calls; the credit cost table at
    https://docs.scrapingant.com/credits-cost applies. Failed requests cost 0 credits.
x-evidence:
  fetched: '2026-08-29'
  method: POST
  url: https://api.scrapingant.com/mcp/
  body: '{"jsonrpc":"2.0","id":1,"method":"tools/list"}'
  http_status: 200
  content_type: text/event-stream
  tools_returned: 3
  credentials_sent: none
maintainers:
- FN: Kin Lane
  email: kin@apievangelist.com