Quickstart

This quickstart walks you from “never connected an MCP server” to “searching your matters from your AI assistant” in under five minutes.

Prerequisites

• A Consilio user account (the same one you use for the portal).
• An MCP-compliant client. Examples include Claude Desktop, GitHub Copilot, or Cursor.

Step 1 — Open your client’s connector settings

Aurora MCP works with any client that supports the Model Context Protocol. If you’re using Claude, you have several options:

• claude.ai — Settings → Connectors → Add custom connector.
• Claude Desktop — Settings → Developer → MCP servers.
• Claude Code — claude mcp add (see Claude Code setup).

Step 2 — Point it at Aurora MCP

Use this endpoint. This same URL works with any compliant MCP client:

https://mcp.ai.consilio.com

Your assistant will discover the rest via RFC 9728 (Protected Resource Metadata) and walk you through the OAuth login against ConsilioId.

Step 3 — Sign in as yourself

A browser window opens for auth.ai.consilio.com. Use your normal Consilio credentials. No extra enrolment. No shared client secret.

Step 4 — Ask your assistant something

Try one of:

• “What active matters do I have for Blackstone?”
• “Show me the workspaces inside A12345.”
• “Search the workspace for memos about the SEC investigation.”
• “Find documents about the settlement letter in matter B204891.”

The client will call the appropriate Aurora MCP tool — as you — and provide the information to your assistant for citation.

Troubleshooting

• 401 unauthorised — token didn’t reach the server. Refresh the connector and re-authorise.
• No matters returned — the user has no visible matters in Consilio. Confirm portal access first.
• Slow first response — the MCP server reranks ambiguous matter searches with an LLM; first calls can take a couple of seconds.