Remote MCP Server

Last updated on 17 May 2026

Overview

JSONata Studio provides a remote Model Context Protocol (MCP) server at https://jsonatastudio.com/mcp.

You can connect this server to MCP-compatible AI clients such as VS Code with GitHub Copilot or Cursor. Once connected, your assistant can evaluate JSONata expressions against JSON data by calling JSONata Studio's hosted evaluation tool.

The remote server uses HTTP transport, so there is no local package to install or background process to run.

Setup

VS Code and GitHub Copilot

Create or update .vscode/mcp.json in your workspace:

{
  "servers": {
    "jsonata-studio": {
      "type": "http",
      "url": "https://jsonatastudio.com/mcp"
    }
  }
}

After saving the file, reload VS Code or use the MCP server controls in Copilot Chat to start the server. Copilot should then be able to discover the JSONata Studio tool.

Cursor

Create or update .cursor/mcp.json in your workspace:

{
  "mcpServers": {
    "jsonata-studio": {
      "url": "https://jsonatastudio.com/mcp"
    }
  }
}

After saving the file, reload Cursor or refresh MCP servers from Cursor's MCP settings. The JSONata Studio server should appear as an available MCP server.

Available tool

evaluate_expression

Evaluates a JSONata expression against provided JSON data.

Inputs:

• jsonata: the JSONata expression to evaluate. This field is required.
• json: the JSON data to evaluate the expression against. This field is optional, but most expressions need it.
• bindings: an optional stringified JavaScript object containing values or functions to bind to the expression.

The tool returns a JSON string containing the evaluation result, execution timing, and any error, warning, or informational messages produced during evaluation.

How to use it

Once your AI client is connected to the MCP server, ask it to evaluate a JSONata expression and include the input JSON in your prompt. For example:

Use the JSONata Studio MCP server to evaluate Account.Order.Product.Name against this JSON:

{
  "Account": {
    "Order": {
      "Product": [{ "Name": "Keyboard" }, { "Name": "Mouse" }]
    }
  }
}

The assistant can call evaluate_expression with the expression and JSON data, then explain the returned result.

You can also ask the assistant to debug a failing expression:

Use the JSONata Studio MCP server to evaluate this expression and explain any error: Account.Order.Product.(Name

Privacy note

JSONata Studio does not store the JSON, JSONata expression, or bindings sent to the remote MCP server. Inputs are used only to evaluate the current tool request and return a result.

For each tool use, JSONata Studio collects operational metadata including the tool name, whether JSON was provided, whether bindings were provided, and the length of the JSONata expression.

Because the server still receives the JSON and expression you provide, avoid sending secrets, credentials, or sensitive production data unless you are comfortable sharing that data with the remote service.