Google Gemini

Google's Gemini tooling speaks the Model Context Protocol. The most direct path is the Gemini CLI, which connects to the Kapptivate MCP server over streamable HTTP with a custom authentication header. The same endpoint and header work anywhere Gemini supports MCP tools.

Gemini CLI

Add the server with one command:

gemini mcp add --transport http kapptivate https://mcp.kapptivate.com/mcp \
  --header "X-Kapptivate-API-Key: $KAPPTIVATE_API_KEY"

The URL above serves app.kapptivate.com. Sign in elsewhere and your endpoint carries the name of your environment. Substitute it everywhere on this page.

Or configure it in ~/.gemini/settings.json (all projects) or .gemini/settings.json (one project):

{
  "mcpServers": {
    "kapptivate": {
      "httpUrl": "https://mcp.kapptivate.com/mcp",
      "headers": {
        "X-Kapptivate-API-Key": "${KAPPTIVATE_API_KEY}"
      }
    }
  }
}

Use httpUrl for the streamable HTTP transport (the url field is for SSE, which the server does not use). Header values support ${VAR} substitution, so keep the key in an environment variable rather than committing it.

The API key carries the same access as your Kapptivate account. Treat it like a password, and use a dedicated service account for shared or automated agents.

At organisation scale

For large deployments, point your Gemini agent tooling (including Vertex AI agent builders and Gemini Enterprise) at the same endpoint:

https://mcp.kapptivate.com/mcp

Check this address against the one for your environment before you roll it out to a fleet of agents. Your endpoint gives the rule.

Send the X-Kapptivate-API-Key header on every request and store the key in your secret manager. Scope one key per environment so you can rotate or revoke it independently.

What's next?

Example prompts

What to ask once Gemini is connected.

Copilot Studio

Connect Microsoft 365 Copilot Studio.