Scrupp

What scrupp MCP server does

The scrupp MCP server exposes Scrupp Jobs API operations as MCP tools for search extraction, lead research, and email workflows. It belongs in the 🔎 Search & Data Extraction category and can turn LinkedIn people-search URLs, Sales Navigator URLs, and Apollo search URLs into person records. Other tools enrich LinkedIn profile URLs, retrieve company information, identify people to contact at a company, find verified work emails, and check the deliverability of existing addresses.

The available tool set depends on how the server is connected. The local package includes the broader set of search, enrichment, company, decision-maker, email, job, and credit tools. The public Streamable HTTP connector exposes email finding, email verification, job status, and credit information. Scrupp's private /sn connector provides the broader API for customers who connect through Claude.

How it works

A local installation runs scrupp-mcp-server through npx and authenticates with SCRUPP_API_KEY. Calls create or retrieve jobs through the Scrupp API. Long-running operations can return a running status and job ID rather than holding the conversation open; the agent can then call scrupp_get_job to obtain status and records. SCRUPP_WAIT_SECONDS sets the maximum blocking period and defaults to 120 seconds.

The scrupp MCP server also supports Streamable HTTP. The included HTTP entry point listens on POST /mcp, provides GET /health, and defaults to port 8787. Each request supplies its own bearer key, allowing one process to serve multiple accounts without sharing request credentials. OAuth is supported for the hosted Claude connection; access tokens are checked before requests and can expire after one hour.

Setup and configuration

For a local MCP client, install or run the npm package with an environment object containing SCRUPP_API_KEY. The README shows an npx -y scrupp-mcp-server configuration. Set SCRUPP_WAIT_SECONDS when the default wait limit does not fit the job sizes you expect.

Claude users can add https://mcp.scrupp.com/sn as a custom connector and sign in to Scrupp. Claude Code users can install the supplied plugin or add a deployed HTTP endpoint with an authorization header. The HTTP implementation also recognizes PORT, MCP_PUBLIC_URL, SCRUPP_AUTH_SERVER_URL, and SCRUPP_API_URL; their documented defaults are 8787, https://mcp.scrupp.com, https://app.scrupp.com, and https://api.scrupp.com/api/v1 respectively.

Tools and capabilities

The scrupp MCP server includes these local tools:

• Export people from Sales Navigator, LinkedIn, or Apollo search URLs.
• Enrich LinkedIn profiles and look up company records.
• Find company decision makers and verified work emails.
• Verify existing email addresses, with verification costing one credit per address.
• Check account credits and retrieve results from running jobs.

The private Claude connector additionally supports building a Sales Navigator search URL from a written audience description without using credits. Apollo search requires a connected account. LinkedIn-related and shared-lead-base operations may use a connected Sales Navigator account or Scrupp's shared account pool, according to the Scrupp API and account configuration.

Limitations and notes

The hosted public connector intentionally has fewer tools than the local server. Its restricted set covers email finding, email verification, job status, and credits; LinkedIn collection, profile enrichment, company lookup, decision-maker lookup, and Apollo export remain local-only on that connector.

Results and operation duration depend on Scrupp jobs, account access, available credits, and the supplied search or profile inputs. The server does not itself decide which connected account or account pool is used. For large exports, callers should handle a running job by polling scrupp_get_job rather than assuming the initial request contains all records.