MCP server
Connect Claude, Cursor or any MCP client to Rankfield's remote MCP server over Streamable HTTP, with your API key as a bearer token.
The server is a thin layer over the HTTP API: each tool calls the same endpoint with your key and returns its JSON.
Connect
- URL:
https://rankfield.dev/mcp - Transport: Streamable HTTP (JSON-RPC over
POST), protocol versions2025-11-25,2025-06-18and2025-03-26. - Auth: the header
Authorization: Bearer <key>. Every call needs it,tools/listincluded.
For a client configured with JSON:
JSON
{
"mcpServers": {
"rankfield": {
"type": "http",
"url": "https://rankfield.dev/mcp",
"headers": { "Authorization": "Bearer YOUR_API_KEY" }
}
}
}With Claude Code:
cURL
claude mcp add --transport http rankfield https://rankfield.dev/mcp --header "Authorization: Bearer YOUR_API_KEY"Opening the URL in a browser shows connection instructions; tools only answer authenticated POST requests.
Tools
| Tool | What it does | Same as |
|---|---|---|
get_site_strength | Site Strength of a domain | Get Site Strength |
get_referring_domains | Count and top 20 referring domains | Get Referring Domains |
list_referring_domains | Every referring domain, with cursor | List All Referring Domains |
list_outbound_domains | Every domain it links to | List Outbound Domains |
get_referrer_profile | Referring domains per Site Strength range | Get Referrer Profile |
start_domain_export | Start a CSV or JSONL export (kind: referrers-export or outbound-export) | Start an export job |
get_export_job | Export progress and download links | Get an export job |
get_dataset_status | The graph release in use | Get Dataset Status |
get_bing_suggestions | Bing's autocomplete suggestions for one keyword, in one of ten markets | Get Bing suggestions |
start_bing_suggest_job | Bing suggestions for up to 10,000 keywords, with a delivery deadline | Start a suggest job |
get_bing_suggest_results | Read that job's keywords, with cursor | Get suggest job results |
get_duckduckgo_suggestions | DuckDuckGo's autocomplete suggestions for one keyword | Get DuckDuckGo suggestions |
start_duckduckgo_suggest_job | DuckDuckGo suggestions for up to 10,000 keywords | Start a suggest job |
get_duckduckgo_suggest_results | Read that job's keywords | Get suggest job results |
get_yahoo_search_predictions | Yahoo's search predictions for one keyword | Get Yahoo Search Predictions |
start_yahoo_search_predictions_job | Yahoo predictions for up to 10,000 keywords | Start a suggest job |
get_yahoo_search_predictions_results | Read that job's keywords | Get suggest job results |
extract_page_links | The links on one live page; waits up to 45 seconds, then returns resultsUrl and the job id | Extract links from a page |
start_link_extraction_job | Links on up to 600 pages | Start a Link Extraction job |
get_link_extraction_results | Read that job's pages | Get Link Extraction results |
scrape_page | One page as Markdown, with its metadata, headings, links, images, structured data, language and AI-use signals; takes scope and include | Scrape a page |
start_scrape_job | Scrape up to 10,000 pages; each item takes scrape_page's arguments except timeout | Start a Scrape job |
get_scrape_results | Read that job's items, with cursor | Get Scrape job results |
bot_view | How search engines, AI crawlers and SEO tools are treated by a page's robots.txt, llms.txt, meta robots and TDM signals | Check how crawlers see a page |
start_bot_view_job | Bot View on up to 10,000 pages | Start a Bot View job |
get_bot_view_results | Read that job's items | Get Bot View job results |
brand_profile | A company's name, logos, icons, colours, fonts and social profiles from its domain | Get a company's brand |
start_brand_job | Brands of up to 10,000 domains | Start a Brand job |
get_brand_results | Read that job's items | Get Brand job results |
map_site | A website's URLs from its sitemaps, feeds, homepage links and a short crawl | Map a website |
start_site_map_job | Map up to 1,000 websites | Start a Site Map job |
get_site_map_results | Read that job's items | Get Site Map job results |
site_audit_start | Start a technical SEO and AEO audit of a website; answers at once with its id | Audit a website |
site_audit_list | Your audits, newest first | Follow an audit |
site_audit_status | An audit's status, live progress and issue counts | Follow an audit |
site_audit_summary | The ten main problems in plain words: start here | Read the results |
site_audit_issues | Every issue found, by group, worst first | Read the results |
site_audit_issue_pages | The pages with one issue | Read the results |
site_audit_pages | Page records, by filter, or one page with its issues and links | Read the results |
site_audit_links | Links, by filter (broken, external, nofollow…) | Read the results |
site_audit_redirects | Redirects and their chains | Read the results |
site_audit_duplicates | Groups of duplicate and near-duplicate pages | Read the results |
site_audit_orphans | Known pages no internal link reaches | Read the results |
site_audit_templates | Issues by URL template | Read the results |
site_audit_sitemaps | The sitemaps read and their checks | Read the results |
site_audit_diff | What changed since an earlier audit of the site | Read the results |
site_audit_export | Every page record, 1,000 at a time | Read the results |
site_audit_page_filters | Ready-made page lists with their counts | Read the results |
site_audit_link_filters | Ready-made link lists with their counts | Read the results |
site_audit_structure | The site as a folder tree, by status, type, depth or indexability | Read the results |
site_audit_log | The crawl log: URLs per minute, speed and the latest fetches | Read the results |
site_audit_history | Every audit of the site, with its health score and counts | Read the results |
site_audit_cancel | Stop an audit; what was found stays readable | Follow an audit |
Each tool takes the same inputs as its endpoint (domain, url or urls, query or queries, cursor, limit, maxAge…), and tools/list gives the exact schemas. The Scrape, Bot View, Brand and Site Map job tools take items: each item is one call's arguments without timeout.
Example prompts
- "What is the Site Strength of example.com, and which of its referring domains are above 60?"
- "List the external links on https://example.com/pricing and flag the nofollow ones."
- "Export every domain linking to example.com as CSV."
- "What does Bing suggest for running shoes in the en-GB market?"
- "Scrape https://example.com/pricing and list the plans and prices."
Limits are those of the endpoints behind each tool: see Rate limits.