MCP Icp Fit Scorer

What mambalabsdev/mcp-icp-fit-scorer does

mambalabsdev/mcp-icp-fit-scorer evaluates a company against an ideal customer profile and produces a flat JSON result suitable for Clay, CRM workflows, or downstream agent processing. The primary input is a company domain, with an optional display name. Results include an icp_score from 0 to 100, an icp_tier from A to D, and a breakdown of the signals that contributed to the result.

Scoring is read-only. The package acts as a client for the Mamba Labs ICP Fit Scorer actor on Apify rather than performing the scoring locally.

How it works

An ICP can be supplied in one of three forms:

• A built-in template such as full_signal, saas_outbound, b2b_services, fintech, smb_local, or enterprise.
• A custom JSON scoring_config containing the scoring weights.
• A plain-language icp_description, interpreted with a caller-provided OpenAI or Anthropic key.

Set fetch_signals to true when the actor should collect hiring and technology-stack information before scoring. The request can also include known signals directly, including GTM hiring, open GTM roles, CRM usage, sales-sequencing tools, technology stack, headcount, employee band, revenue estimate, and funding status.

Tier boundaries default to A at 80 or higher, B at 60 or higher, C at 40 or higher, and D below 40. tier_thresholds can replace those boundaries. Other options support funding recency, minimum output scores, previous-score comparisons, and explanatory text.

Setup and configuration

Running the package requires Node.js 18 or newer and an Apify account with an API token. For a standard local MCP configuration, launch the package through npx and set APIFY_TOKEN in the server environment:

{
  "mcpServers": {
    "mamba-icp-scorer": {
      "command": "npx",
      "args": ["-y", "@mambalabsdev/mcp-icp-fit-scorer"],
      "env": {
        "APIFY_TOKEN": "your-apify-token"
      }
    }
  }
}

An llm_api_key is additionally required when using icp_description; llm_provider selects openai or anthropic. Each scoring request uses Apify and consumes Apify credits.

Tools and capabilities

The server exposes score_icp_fit for company evaluation. Its capabilities include:

• Applying predefined or custom weighted ICP criteria.
• Returning a flat, Clay-compatible JSON row.
• Collecting hiring and technology-stack signals through Apify.
• Adding a score explanation when requested.
• Reporting score changes and trends when a previous score is supplied.
• Filtering output below a configured minimum score.

The package documentation also describes a list_icp_templates tool for discovering available templates.

Limitations and notes

The server depends on Apify for its scoring work, so an Apify token and available credits are required. Plain-language ICP parsing adds a dependency on a caller-supplied OpenAI or Anthropic API key. The output is a read-only score and does not describe an operation for modifying company records or pushing data to Clay or a CRM.