Rube is a Model Context Protocol (MCP) server that connects your AI tools to 500+ apps like Gmail, Slack, GitHub, and Notion. Simply install it in your AI client, authenticate once with your apps, and start asking your AI to perform real actions like "Send an email" or "Create a task."
[ MCP SERVER SUMMARY ]
Rube is a Model Context Protocol (MCP) server that connects your AI tools to 500+ apps like Gmail, Slack, GitHub, and Notion. Simply install it in your AI client, authenticate once with your apps, and start asking your AI to perform real actions like "Send an email" or "Create a task." Connect Rube through Bifrost MCP Gateway to centralize access control, virtual keys, and tool-call observability for your AI assistants.
| Property | Details |
|---|---|
| Company | Upstream maintainer |
| Connection type | Confirm upstream transport |
| Authentication | Varies by upstream server |
| Package or endpoint | See upstream documentation |
| Capabilities | Rube tools over MCP, AI assistant connectivity, Governed tool access through Bifrost |
| Official documentation | Rube MCP documentation |
[ HOW TO CONNECT ]
Start the gateway locally or deploy in your VPC.
npx -y @maximhq/bifrost
Open MCP Gateway in the Bifrost dashboard, click New MCP Server, and select the transport documented by the upstream server. Or register via the API below.
• Confirm the transport, command or endpoint, and authentication requirements in the Rube upstream documentation.
• Replace every placeholder in the example before sending the request.
Configure Claude Desktop, Cursor, or Claude Code to use your Bifrost endpoint with a virtual key instead of connecting directly to Rube.
{
"mcpServers": {
"bifrost": {
"url": "http://localhost:8080/mcp",
"headers": {
"Authorization": "Bearer vk_your_virtual_key"
}
}
}
}[ EXAMPLE CONFIG ]
Add this MCP client with a POST to /api/mcp/client. See Connecting to MCP Servers for full options.
# For a remote HTTP or SSE server:
curl -X POST http://localhost:8080/api/mcp/client \
-H "Content-Type: application/json" \
-d '{
"name": "rube",
"connection_type": "<http|sse>",
"connection_string": "<upstream MCP endpoint>",
"tools_to_execute": ["*"]
}'
# For a local STDIO server:
curl -X POST http://localhost:8080/api/mcp/client \
-H "Content-Type: application/json" \
-d '{
"name": "rube",
"connection_type": "stdio",
"stdio_config": {
"command": "<upstream command>",
"args": ["<arg>"]
},
"tools_to_execute": ["*"]
}'[ WHY BIFROST ]
Virtual keys, RBAC, and tool allowlists per team or agent.
Attribute MCP tool spend alongside LLM usage in one view.
Log every tool call with latency, server, and virtual key metadata.
Upstream secrets stay in the gateway — not on every laptop.
[ FAQ ]
Rube is a Model Context Protocol (MCP) server that connects your AI tools to 500+ apps like Gmail, Slack, GitHub, and Notion. Simply install it in your AI client, authenticate once with your apps, and start asking your AI to perform real actions like "Send an email" or "Create a task." The exact tools available depend on the upstream server version and configuration.
Check the upstream documentation to determine whether Rube uses STDIO, Streamable HTTP, or SSE. Then add it as an MCP client in Bifrost using the corresponding command or endpoint.
No. Bifrost connects to Rube as an upstream MCP client and adds a governed gateway between that server and your AI assistants.
[ RELATED SERVERS ]

Improve and quality control your outputs with evaluations using LLM-as-Judge
Unified integration layer that bridges multiple AI coding assistants (Codex, Claude Code, Cursor, Gemini) through zero-configuration auto-discovery and enterprise-ready architecture.
MCP server to interact with Routine: calendars, tasks, notes, etc.