Skip to main content

Quick Start

Get up and running with MCP Gateway in 4 steps.

Add MCP
Source: MCP Library
Server: GitHub MCP
Transport: SSE
Configure
Tools: 12 enabled
Guardrails: PII detection ON
Access: Cursor, Claude
Connect Agent
url: mcp.quilr.ai/github/mcp
auth: Bearer <token>
mcpuser: user@company.com
Monitor
Tool calls: 3,421
Agents: 4 connected
Blocked: 12 requests
QuilrAI

1. Add an MCP Server

Go to the MCP Gateway tab and install from the MCP Library for one-click pre-built integrations, or click Add MCP to register any server by its transport URL (ending in /sse or /mcp).

OAuth MCPs will prompt you to authorize; token-based MCPs receive API tokens automatically.

2. Configure Your MCP

Open Settings on any MCP card to fine-tune its behavior. Sensible defaults are applied automatically.

SettingDescription
Tools ManagementEnable/disable tools by category (read-only, write, destructive)
Security GuardrailsPII/PHI/PCI detection, adversarial blocking
Access ControlRestrict which AI agents can use this MCP
Web Search PolicyDomain exclusions via firewall integrations
OAuth ConnectOne-click OAuth authorization flow
Agents ConfigurationDefine and manage AI agent profiles

3. Connect Your Agent

Point your AI agent or client to the MCP endpoint URL shown on the card. Use Authorization: Bearer <token> and mcpuser headers for token-based MCPs. OAuth MCPs use the Connect flow instead.

{
"mcpServers": {
"quilr-mcp": {
"url": "https://mcp.quilr.ai/<your-mcp-id>/mcp",
"headers": {
"Authorization": "Bearer <your-token>",
"mcpuser": "user@company.com"
}
}
}
}

See the Integration Guide for more client examples.

4. Monitor Tool Calls

Every tool call through the gateway is logged with tool name, parameters, guardrail actions, and user identity. Check your Logs tab to verify requests are flowing through.