Data Orlando

What Data Orlando MCP server does

Data Orlando exposes Orlando Open Data, the City of Orlando’s Socrata catalog, through MCP. The main workflow is to find a dataset with datasets, inspect it with metadata, and then retrieve records with query.

The catalog search accepts a keyword and returns dataset identifiers, names, descriptions, categories, and update dates. The identifier, called a resource_id, is used by the other tools. Metadata responses include the dataset schema, column names and types, row count, category, and last-updated information.

query runs a Socrata SoQL request against a selected dataset and returns matching rows as JSON. It accepts where, select, group, and order clauses without the leading $, along with limit and offset for result sizing and pagination.

How it works

The hosted MCP endpoint is:

https://gateway.pipeworx.io/data-orlando/mcp

An MCP client connects to that URL and discovers the Data Orlando tools. The endpoint also exposes shared Pipeworx gateway tools, including ask_pipeworx, discover_tools, search_within, and memory-related tools. Those shared tools can route broader questions across Pipeworx data sources; they are separate from the three Data Orlando tools described above.

The full Pipeworx endpoint, https://gateway.pipeworx.io/mcp, is another way to reach the same gateway and data sources. A scoped Data Orlando connection lists the pack’s tools directly while still having access to gateway routing through ask_pipeworx.

For clients or scripts that do not use MCP, individual tools can also be called through the gateway’s HTTP API. For example, the dataset search endpoint is POST https://gateway.pipeworx.io/v1/tools/data_orlando_datasets with a JSON body containing a query field.

Setup and configuration

Add the hosted endpoint to an MCP client such as Claude Desktop, Cursor, or Windsurf:

{
  "mcpServers": {
    "data-orlando": {
      "url": "https://gateway.pipeworx.io/data-orlando/mcp"
    }
  }
}

A local stdio bridge is available through npm. The bridge forwards calls to the hosted gateway, so it requires network access and has no offline mode:

{
  "mcpServers": {
    "data-orlando": {
      "command": "npx",
      "args": ["-y", "@pipeworx/mcp-data-orlando"]
    }
  }
}

Anonymous use is available with a daily call limit. A free Pipeworx account increases the limit, and its API key can be supplied as PIPEWORX_API_KEY in the MCP server’s environment. The key is sent as a Bearer token.

The npm bridge lists the Data Orlando tools by default. Set PIPEWORX_META_TOOLS=1 to include the shared gateway tools as well. When using an HTTPS proxy, newer Node.js releases can honor the proxy environment settings when NODE_USE_ENV_PROXY=1 is enabled.

Tools and capabilities

• datasets: Search Orlando’s open-data catalog by keyword and return identifiers and descriptive dataset details.
• metadata: Inspect a dataset’s fields, types, row count, category, and update timestamp.
• query: Execute SoQL selections, filters, grouping, ordering, limits, and offsets against a dataset.
• ask_pipeworx: Ask a plain-language question through the gateway’s cross-source routing tool when shared meta-tools are enabled or when using the hosted gateway.

The package is published as @pipeworx/mcp-data-orlando and is licensed under MIT.

Limitations and notes

The npm option is a network bridge to the hosted gateway, not a local copy of Orlando’s data. Calls therefore depend on connectivity to Pipeworx and the upstream data service. Anonymous calls are capped at 50 per day; a free account raises that allowance to 200 per day. Hosted and npm connections use the same gateway, while the tools shown directly can differ by endpoint configuration.