Skip to main content

​
Connecting Liberator to Microsoft Copilot Studio

Add Liberator as an MCP connector in a Copilot Studio agent using OAuth 2.0, so each user signs in with their own Liberator credentials. Configure this from within Copilot Studio (not Microsoft 365 or Teams). An existing agent is required.

​
Prerequisites

  • Access to Microsoft Copilot Studio with permission to edit an agent
  • An existing Copilot Studio agent
  • A CloudQuant Data Liberator account with valid username and password
  • The Liberator MCP Endpoint URL from the Liberator UI (for example, https://api.cloudquant.ai/mcp)
Requires Streamable HTTP transport and OAuth 2.0 Dynamic Client Registration (DCR). The connector can be shared org-wide, but each user creates their own connection by signing in with their own Liberator credentials. No shared API keys are needed.

​
Setup steps

1

Add an MCP tool

In your agent, go to Tools → New Tool and select Model Context Protocol (MCP).
2

Enter server details

Enter a server name and description, and set Server URL to your Liberator MCP Endpoint URL, for example:
3

Configure OAuth 2.0

Under Authentication, select OAuth 2.0 with Type set to Dynamic discovery, then click Create.
4

Create a connection

On the Add tool screen, open the Connection dropdown, select Create new connection, then Create.
5

Sign in to Liberator

Sign in with your Liberator username and password and grant access (profile, openid, email).
6

Add and configure

Click Add and configure. Copilot auto-registers the Liberator tools (list_datasets, get_dataset_config, query_dataset).

​
Using the connector

After the tool is added, Liberator MCP tools are available to the agent: Each user who uses the connector signs in with their own Liberator credentials, so entitlements stay per-user.

ChatGPT

Connect ChatGPT with the MCP connector and OAuth

Claude Desktop

Install the Liberator .mcpb extension in Claude Desktop