Google Gemini CLI

Google’s gemini-cli supports MCP servers through ~/.gemini/settings.json (user-scoped) or .gemini/settings.json in a project root (project-scoped).

Setup

1. Get a Bearer token by visiting https://<your-workspace>.plaid.cloud/mcp/setup/token in a browser where you’re signed into PlaidCloud.

2. Edit ~/.gemini/settings.json (create it if it doesn’t exist):

   {
     "mcpServers": {
       "plaidcloud": {
         "httpUrl": "https://<your-workspace>.plaid.cloud/mcp/",
         "headers": {
           "Authorization": "Bearer eyJhbGc…"
         }
       }
     }
   }

3. Restart gemini. Inside a session, run /mcp to verify the server is connected and list its tools.

Usage

Prompt Gemini to use the PlaidCloud tools directly:

• “Use plaidcloud to list workflows in the Sales Forecasting project.”
• “Describe the columns of the customers table and the most recent snapshot.”

You can also call tools explicitly with the /tools command if you want to inspect a specific tool’s input schema before invoking it.

Refreshing the Token

When the token expires, reload https://<your-workspace>.plaid.cloud/mcp/setup/token and update the Authorization value in settings.json. Restart gemini to pick up the change.

Gemini Code Assist (IDE)

Gemini Code Assist in VSCode/JetBrains accepts the same mcpServers config under its agent settings. The schema matches the CLI form — paste the same snippet under the IDE’s MCP server settings panel.