Connecting to the Regrid MCP Server

The Regrid MCP Server is available at:

https://support.regrid.com/mcp

MCP client support varies by platform — some use a point-and-click UI, others require editing a local JSON config file. This guide covers current setup steps for each major client as of this writing.

Quick reference

Claude (claude.ai, Claude Desktop, Cowork, mobile)

This is the connection method most teams are already familiar with, since it's brokered through your Anthropic account rather than a local device — meaning it works identically across web, desktop, Cowork, and mobile once configured.

Prerequisites

• A Pro, Max, Team, or Enterprise plan (custom connectors aren't available on the Free plan)
• The MCP server must be reachable over the public internet — Claude's cloud infrastructure cannot reach servers only exposed on a VPN or internal network

Steps

1. Click the + button next to the chat input (web/desktop) or the paperclip icon (mobile).
2. Select Connectors → Add custom connector. (Alternatively: Settings → Connectors → Add custom connector.)
3. Enter the connector details:

   Field	Value
   Name	Regrid MCP Server
   URL	https://support.regrid.com/mcp

4. Click Connect. If the server requires authentication, Claude will walk you through an OAuth consent flow.
5. Once added, the connector appears in the Connectors menu (paperclip/"+" icon), where you can toggle it on or off per conversation.

Verifying the connection
Start a new chat, toggle the connector on, and ask something like:

"Use the Regrid MCP to list the available Regrid API endpoints."

Gemini

Gemini web/mobile app (Gemini Spark)

Google refers to this integration as a Connected App.

Prerequisites

• Access to Gemini Spark
• A personal Google Account — Workspace and school (Education) accounts are not currently supported
• "Keep Activity" turned on in your Gemini settings

Steps

1. Go to gemini.google.com.
2. Open Settings & help → Connected Apps.
3. Select Add a custom app and enter the MCP server URL: https://support.regrid.com/mcp.
4. Complete whatever authentication the server requests (OAuth consent screen, if applicable).
5. Enable the connected app within a conversation to start using it.

Gemini Enterprise (Google Cloud)

If your organization uses Gemini Enterprise rather than the consumer Gemini app, setup happens through the Google Cloud Console instead:

1. In the Google Cloud Console, navigate to Data Stores → Custom MCP Server.
2. Register the server URL and select Streamable HTTP as the transport (the deprecated SSE transport is not supported here).
3. Configure access permissions per your organization's IAM policy.

Gemini CLI

For teams using the Gemini CLI, MCP servers are configured through a local settings.json file rather than any UI:

{
  "mcpServers": {
    "regrid-support": {
      "url": "https://support.regrid.com/mcp"
    }
  }
}

ChatGPT / OpenAI

OpenAI calls custom MCP connections part of Developer Mode.

Prerequisites

• A Plus or Pro plan (Developer Mode / full custom MCP tool support is not available on Free)
• Note: earlier iterations of this feature restricted custom connectors to Deep Research mode only. With Developer Mode enabled, connectors are now also available in regular chat.

Steps

1. Go to Settings → Advanced and enable Developer Mode.
2. Start a new chat and click Add sources / Add connector near the + icon on the message composer.
3. Add a new connector using the MCP server URL: https://support.regrid.com/mcp.
4. Complete authentication if prompted (OAuth or API key, depending on server configuration).
5. To use the connector in a conversation, toggle it on and reference it explicitly in your prompt, e.g.:

   "Use the Regrid Support tool to look up parcel data for this address."

Cursor

Cursor stores MCP configuration in a JSON file named mcp.json.

Steps

1. Open Cursor Settings → Tools & MCP.
2. Click New MCP Server (this opens mcp.json for editing), or create the file manually at one of:
   • .cursor/mcp.json — project-scoped; check this into version control to share the connector with your team
   • ~/.cursor/mcp.json — global; available across every project on your machine
3. Add an entry:

{
  "mcpServers": {
    "regrid-support": {
      "url": "https://support.regrid.com/mcp"
    }
  }
}

4. Save the file, then restart Cursor (or reload the MCP panel from the Tools & MCP settings page).
5. If the server requires OAuth, Cursor will automatically trigger a browser-based authentication flow and securely store the resulting credentials.

GitHub Copilot (VS Code)

Copilot's MCP support lives in VS Code's mcp.json and requires Agent mode.

Prerequisites

• VS Code 1.101+ for reliable remote/HTTP MCP and OAuth support

Steps

1. Open the Command Palette (Cmd/Ctrl+Shift+P) and run:
   • MCP: Open User Configuration — for a config available across every workspace, or
   • Create .vscode/mcp.json manually for a project-scoped config.
2. Add a server entry:

{
  "servers": {
    "regrid-support": {
      "type": "http",
      "url": "https://support.regrid.com/mcp"
    }
  }
}

3. Save the file — VS Code will prompt you to start the server. Run MCP: List Servers from the Command Palette to confirm it connected successfully.
4. Open Copilot Chat and switch the mode dropdown from Ask to Agent — MCP tools are invisible in any other mode.
5. Copilot will ask for confirmation the first time it invokes a tool from the server.

Perplexity

Perplexity calls this a Custom Remote Connector, available on Pro and Enterprise plans.

Steps

1. Go to Account Settings → Connectors (for org-wide setup on Enterprise: Enterprise Settings → Permissions → Connectors permissions).
2. Click + Custom connector, then select Remote.
3. Fill in the connector details:

   Field	Value
   Name	Regrid Support MCP
   MCP Server URL	https://support.regrid.com/mcp
   Transport	Streamable HTTP
   Authentication	None / API Key / OAuth 2.0 — depending on how the server is configured

4. Check the risk-acknowledgement box and click Add.
5. Click the new connector card to complete authentication and enable it, then toggle it on within a chat thread to use it.

Troubleshooting checklist

• Connector added but tools aren't showing up — confirm the connector/toggle is enabled for the current conversation or session, not just added at the account level.
• mcpServers vs. servers — double-check the config root key matches the platform (Cursor uses mcpServers; VS Code/Copilot uses servers).
• Auth errors after adding a header-based token — re-verify the header value doesn't have stray whitespace or line breaks, and that any environment-variable placeholders (e.g. ${REGRID_TOKEN}) are actually set in the shell or config's env block.
• SSE vs. Streamable HTTP — if a platform lets you pick a transport explicitly, choose Streamable HTTP.
• Still stuck? Check the platform's own current MCP/connector documentation, since UI names and requirements shift frequently across all these products.