Agent to Brokerage Resolver MCP

What Agent to Brokerage Resolver MCP does

Agent to Brokerage Resolver MCP connects an MCP client to the Mamba Labs Agent to Brokerage Resolver actor on Apify. It checks licensed US real estate agents against state license-register data and can return the brokerage of record, the agent's license status, and the license expiry date.

The resolver also reports the brokerage's active-agent count and places it in one of four size bands: 1, 2–15, 16–50, or 51+. In roster modes, it can return active agents at a brokerage or brokerage offices located in a ZIP code. The project is listed in the 🏠 Real Estate category and is intended for workflows that need brokerage or license information rather than personal contact data.

Coverage spans 19 states, while brokerage data is available in California, Iowa, Missouri, Nevada, New York, Ohio, Oregon, Texas, and Vermont. ZIP-code office searches are available for New York, Missouri, Oregon, and Vermont.

How it works

The package is a thin MCP client for an Apify actor. A tool call starts an actor run, waits for completion, and returns the resulting dataset rows. The actor reads Mamba Labs' index of state license registers through its service.

The single tool is resolve_agent_brokerage. Its mode can be lookup, roster_firm, or roster_zip, with lookup used by default. Lookup requests can identify an agent by license number or by first and last name. Name searches can be narrowed with a brokerage, city, or license county. Batch lookups accept newline-separated license or agent records.

roster_firm accepts a brokerage name or brokerage license number and returns its agents. roster_zip accepts a five-digit ZIP code and can optionally include each office's active agents. Roster calls support an active-only setting and a maximum row count. A maximum charge can also stop a run before it exceeds a specified amount.

Ambiguous, unmatched, unsupported-state, and capped inputs produce rows with a status explanation and are not charged. A failed run is returned as an error with its run ID.

Setup and configuration

Run the server with Node.js 18 or newer and an Apify account. Configure it as a stdio MCP server using npx and the package @mambalabsdev/mcp-agent-brokerage-resolver. The required environment variable is APIFY_TOKEN, containing an Apify API token.

A minimal client configuration uses the following structure:

{
  "mcpServers": {
    "mamba-agent-brokerage-resolver": {
      "command": "npx",
      "args": ["-y", "@mambalabsdev/mcp-agent-brokerage-resolver"],
      "env": {"APIFY_TOKEN": "your-apify-token"}
    }
  }
}

The server can list its tool without a token, but a tool call requires one. The actor itself does not require a provider key. Apify charges the account by event: matched lookup rows, roster rows, and actor starts have separate rates by plan.

Tools and capabilities

Agent to Brokerage Resolver MCP exposes one tool with these capabilities:

• Look up one agent by state and license number.
• Find an agent by name, with optional brokerage or city filters.
• Process batches of license numbers or agent records.
• List active agents associated with a brokerage.
• List brokerage offices in a ZIP code.
• Include office agents in ZIP-code roster results.
• Limit roster rows and set an optional run-charge ceiling.
• Return license status, expiry, brokerage, active-agent count, and size band where supported.

The output does not include emails, phone numbers, or home addresses.

Limitations and notes

Agent to Brokerage Resolver MCP depends on the states and data types supported by the underlying actor. State coverage is not uniform: brokerage information is explicitly available only for the nine states listed above, and ZIP-code office rosters are limited to four states.

License numbers should be supplied in the format used by the relevant register. Name matching can be ambiguous or produce no match, in which case the returned row explains the result. Roster result limits range from 1 to 1,000 rows per brokerage or ZIP code, with a default of 100.

Apify usage is billed to the configured Apify account. The service checks the run token against the Apify API for each request and does not cache results or store or log the token, according to the project documentation. The project is released under the MIT license.