MCP Review Platform Reputation Enricher

What mcp-review-platform-reputation-enricher does

The mcp-review-platform-reputation-enricher MCP server turns a company domain or Trustpilot business unit ID into one flat JSON record. The result includes the Trustpilot rating, total review count, star score, one-to-five-star distribution, Trustpilot URL, country code, and information about how the business was found. An optional company name is carried into the output for joining with other records, but it does not affect the lookup.

The primary tool is get_trustpilot_reputation. It returns 33 snake_case fields rather than nested objects, which suits downstream enrichment and tabular workflows. Consumers should inspect trustpilot_status before using rating fields. The response also includes rating_is_meaningful, based on a configurable minimum review count, so a high rating with very few reviews can be distinguished from a rating supported by a larger sample.

How it works

The server is a thin client for the Mamba Labs Trustpilot Reputation Enricher actor on Apify. With the default route, it can inspect a company's own Trustpilot widget embed to find a business unit ID, then read Trustpilot's public TrustBox data endpoint. Supplying the 24-character trustpilotBusinessUnitId is more reliable and bypasses domain-based ID discovery.

Set publicPageAccess to unblocker to use the Trustpilot public company page through Apify Unblocker instead. This route can resolve a domain without a known business unit ID and can add categories, claimed status, and verification data. The server passes the actor result back without implementing separate lookup logic locally.

Successful lookups are cached for seven days by default. Set skipCache to true when a fresh fetch is required. The minReviewCount setting accepts none, 10, 50, 100, or 500; it changes the meaningfulness flag but does not remove records or alter ratings.

Setup and configuration

Run the package with Node.js 18 or newer, an Apify account, and an Apify API token. A Claude Desktop stdio configuration can launch it through npx:

{
  "mcpServers": {
    "mamba-review-platform-reputation-enricher": {
      "command": "npx",
      "args": ["-y", "@mambalabsdev/mcp-review-platform-reputation-enricher"],
      "env": {
        "APIFY_TOKEN": "your-apify-token"
      }
    }
  }
}

The mcp-review-platform-reputation-enricher MCP server requires APIFY_TOKEN; no Trustpilot API key is needed. Restart the client after adding the configuration so the tool becomes available.

Tools and capabilities

The mcp-review-platform-reputation-enricher MCP server accepts these relevant inputs:

• company_domain, an optional bare domain such as monzo.com.
• trustpilotBusinessUnitId, the preferred and most dependable identifier.
• company_name, copied into the returned row for record matching.
• publicPageAccess, set to off or unblocker.
• minReviewCount, used to calculate rating_is_meaningful.
• sources, currently limited to Trustpilot.
• skipCache, used to force a new lookup.

The output distinguishes blocked, not_found, not_extractable, and skipped states. G2, Capterra, and Glassdoor are represented as skipped columns in the current version, so the response shape remains stable while those sources are unavailable.

Limitations and notes

A domain-only lookup succeeds when the company's own site exposes a Trustpilot widget with a business unit ID. Many domains therefore return not_found unless an ID is supplied. The public-page route is an alternative, but it consumes Unblocker units from the user's Apify account and accesses Trustpilot pages that Trustpilot's robots.txt disallows to automated agents.

Claimed status and verification are available through the public-page route; the default route may report claimed status as not_extractable. The includeCategories option has no effect on the default route, and categories remain null unless publicPageAccess is unblocker. The current source implementation queries Trustpilot only, while other review-platform columns remain explicitly skipped.