PortyLink is a remote MCP server. Any MCP client that supports remote servers with OAuth can connect to it. You approve each agent once, choosing the organization, the sites and the data it may read; it can only read, never change anything.
The server address is:
https://portylink.com/mcpA connector added in Claude Desktop, claude.ai or the mobile app is available in all three for your Claude account.
If your plan doesn't offer custom connectors, use the bridge below. In Claude Desktop, open Settings → Developer → Edit Config, add this to claude_desktop_config.json and restart Claude. It needs Node.js; a browser window opens for you to sign in.
{
"mcpServers": {
"portylink": {
"command": "npx",
"args": ["-y", "mcp-remote", "https://portylink.com/mcp"]
}
}
}claude mcp add --transport http portylink https://portylink.com/mcpThen run /mcp in Claude Code to sign in.
In Settings → Apps & Connectors → Advanced, turn on Developer mode, then create a connector with the address above and OAuth authentication.
Add this to ~/.cursor/mcp.json (or the project's .cursor/mcp.json):
{
"mcpServers": {
"portylink": { "url": "https://portylink.com/mcp" }
}
}| Tool | What it does |
|---|---|
list_sites | The sites the agent may read, with which providers each has linked. |
get_report | Runs one of the reports for a site and date range. |
compare_periods | A report for a period and for the period before it, or the same period a year earlier, with the changes. |
Ask in plain language, for example:
Every connected agent appears on your organization's AI agents page with what it may read and when it was last used. Disconnect cuts it off on its next request. Only owners and admins can connect agents.
Agents are limited to 120 requests a minute each. Data from the last two to three days is provisional in Search Console; reports mark it.