Atlassian MCP / App in amberAI for Users

Control Jira & Confluence directly from amberAI – Guide for End Users

What is the Atlassian MCP Server?

MCP (Model Context Protocol) servers are an abstraction layer for integrating external tools and capabilities into amberAI. amberAI already includes many built-in tools that let you search your internal documents, browse the web, generate images, and much more. However, these tools do not interact with external applications such as Jira or Confluence.

This is where the Atlassian MCP Server comes in: it enables you to perform actions in Jira and Confluence directly from an amberAI conversation for example, creating issues, managing sprints, or building Confluence pages.

💡 The Atlassian MCP Server supports both Atlassian Cloud and On-Premise / Data Center.

Setup & Connection

Prerequisites

· An active amber account with access to amberAI

· An Atlassian account (Cloud or Data Center)

· The Atlassian MCP Server has been set up by your administrator

How to connect

Step 1: Open user settings via your Name on the bottom left in amber

Navigate to User Settings → Connectors → Apps.

Step 2: Find the Atlassian MCP Server

In the app overview you will see the Atlassian MCP Server provided by your administrator.

Step 3: Establish the connection

Click on the MCP Server and follow the authentication flow (e.g. OAuth or API token). This step is required to retrieve all tools and functions and make them available in amberAI.

Step 4: Reload the page

Reload the amber page once – the Atlassian integration is now active.

🔒 Each user must connect to the MCP Server individually. amber always respects your personal access permissions.

Available Actions

Jira Actions

Category	Action	Description
🔍 Search	Search issues	Jira search via JQL; filter issues by project, status, or label
📖 Read	Get issue details	View status, priority, description, assignments, and comments
➕ Create	Create new issue	Specify issue type, priority, labels, description, and assignee
✏️ Edit	Update issue	Change status, update fields, assign or re-prioritize issues
🗑️ Delete	Delete issue	Permanently delete an issue (only with appropriate permissions)
💬 Comments	Manage comments	Read, create, and edit comments
🏃 Sprints	Sprint management	Create, start, and complete sprints (Scrum)
📋 Boards	Board management	View boards and their configurations
📁 Projects	Project overview	View project details and configurations

Confluence Actions

Category	Action	Description
🔍 Search	Search pages	Confluence search (CQL); filter pages by space
📖 Read	Get page content	Read content, metadata, and comments
➕ Create	Create new page	Create a Confluence page with title, content, and space
✏️ Edit	Update page	Modify and extend existing page content
🗂️ Navigation	Page hierarchy	Navigate sub-pages and space structures

Managing Permissions

You can define how amberAI should handle each individual tool of the Atlassian MCP Server. Open the MCP Server in your user settings and adjust the permissions per tool:

Setting	Behavior
Always allow	amberAI uses the tool automatically without asking you first.
Always ask	You receive a notification in the conversation and can approve or decline the action. (Default setting)
Always deny	The tool will not be used in the conversation.

⚠️ Be cautious when setting permissions to "Always allow" – especially for tools that perform write operations (e.g. creating issues or editing pages). This may lead to unintended changes if amberAI uses the tool without your explicit approval.

Using the MCP Server in amberAI

Activating the MCP Server

1. Open a new or existing conversation in amberAI.

2. Click on "Apps" in the options area below the chat input.

3. Enable the Atlassian MCP Server to make its tools available for the conversation.

4. Enter your request in natural language.

How it works in the chat

When amberAI wants to use a tool from the Atlassian MCP Server, it checks your permissions:

· With "Always ask": You receive a notification and can approve or decline.

· With "Always allow": amberAI uses the tool directly and shows you the results.

· With "Always deny": amberAI skips the tool and acts accordingly.

Our recommendation is to set the mode to "Always allow"; this will save you time with every MCP request by reducing the number of authorization requests from amberAI.

Example prompts

You can ask amberAI to perform Atlassian actions in natural language:

· "Create a Jira ticket in the MARKETING project with high priority and assign it to John Smith."

· "Show me all open issues in the current sprint of the WEBDEV project."

· "Create a new Confluence page in the Documentation space titled Meeting Minutes."

· "Change the status of ticket ABC-123 to In Progress."

· "Summarize all comments on issue DEF-456."

· "Which issues are currently assigned to me in the BACKEND project?"

· "Create a new sprint on the Engineering board and add the open bugs."

Recommendation: Search vs. MCP Actions

Purpose	Recommended Tool	Why?
Finding knowledge (searching tickets, pages, comments)	amberSearch	Index-based, semantic search across Jira & Confluence with permission awareness and source references.
Performing actions (creating issues, changing status, building pages)	Atlassian MCP Server	Direct interaction with Jira & Confluence from amberAI via the Model Context Protocol.

💡 Best Practice: Combine both approaches! Use amberSearch for fast knowledge retrieval and the MCP Server for follow-up actions directly within the conversation.

Frequently Asked Questions (FAQ)

Q: Do I need to connect separately even though my admin has already set up the MCP Server?

A: Yes. Each user must connect individually via user settings in order to use the tools in their conversations.

Q: Can I add the MCP Server myself as a developer?

A: Yes. Users with the Administrator and Developer role can add their own MCP servers under User Settings → Connectors → Apps. These are visible and usable only by you.

Q: What happens if I decline a tool?

A: amberAI will not use the declined tool and will act accordingly in the conversation – it will look for alternative approaches or inform you.

Q: Are my Atlassian permissions respected?

A: Yes. The MCP Server operates with your personal credentials. You can only access data and perform actions for which you are authorized in Jira/Confluence.

Availability

The Outlook MCP is currently available as a Beta feature. This means the integration is already usable, but some parts may still be limited, refined, or changed over time as the feature continues to improve.

amber Help Center | Atlassian MCP Server – Guide for End Users