Configure tools for an external MCP server
An MCP server is a service that implements the Model Context Protocol (MCP) to expose tools and data to AI agents in a standardized way. It acts as a bridge between AI clients (like ChatGPT, Claude, or Rovo) and external systems so the AI can securely discover and call those systems’ capabilities (for example, via APIs) using natural language.
The Atlassian Rovo MCP server allows tools, like AI assistants and developer environments, to securely access and interact with your Atlassian data. By connecting Atlassian’s MCP server, you can bring the context of your Jira issues, Confluence pages, and Teamwork Graph directly into your preferred AI development tools and chat interfaces.
Before you begin
Make sure:
You’re an organization admin having permissions to manage MCP servers in Atlassian Administration.
You’ve confirmed with relevant teams (for example, security, platform, or app owners) that it’s safe to configure this MCP server, and that impacted AI use cases have been reviewed.
Configure external MCP server settings
After adding an external MCP server from Atlassian Administration, you can configure which tools can be integrated into Rovo and other AI features by users in your organization.
To configure the settings of an external MCP server:
Go to Atlassian Administration. Select your organization if you have more than one.
From the sidebar menu, expand Apps.
Expand Sites then select the site where the MCP server is located.
Select Connected apps.
From the list of connected apps, locate the MCP server.
Under the Actions column, select View app details.
Select the Tools selection tab.
Find one or more MCP tools you want to deactivate and deselect the checkbox for the tools.
Once deactivated, the tools will no longer be available to users. You can return to the same screen at any time to reactivate tools if needed.
How Rovo credit usage works with MCP
When an AI client calls via MCP to retrieve data or generate insights, it consumes Rovo credits. These credits come from the same shared pool used by Rovo Chat, Studio, Agents, and the Teamwork Graph, across your organization.
Credit usage is based on the complexity of the request:
Context volume: Larger amounts of data pulled from your site require more credits.
Reasoning depth: Complex queries that require more processing power consume more credits than lightweight lookups.
Usage cap: To help you manage costs, credit usage is per call.
For thresholds, plan-level allowances, and what happens when you approach your usage limit, see Rovo usage allowance
Enriched Teamwork Graph (TWG) API and tool calls, including unified search and context tools called via Rovo MCP, consume Rovo credits. These include but are not limited to: getTeamworkGraphContext,twg rovo search, searchAtlassian, search.
Find out how Rovo credits work for enriched TWG API and tool calls.
Was this helpful?