Integrating the Jira MCP server with a local MCP client

[omri-madar]

Hey,

Pretty new to MCP, so maybe it's a newbie question :-)
In the README it says "MCP Atlassian is designed to be used with AI assistants through IDE integration."
Is that the only way to use it? How do I use it if I have a local MCP client? I've set up the MCP client from Anthropic tutorial:
https://modelcontextprotocol.io/quickstart/client

Now I basically should be able to give it an MCP server URL and the client should know how to integrate...
Can I simply run the MCP server as docker "standalone" and my client could work with it?

[sooperset]

Hi @omri-madar,

Yes, you can use the mcp-atlassian server with any MCP client. You can run the server as a standalone Docker container and connect your local MCP client to it.

For a working example, check out the implementation in the SDK repo: https://github.com/modelcontextprotocol/python-sdk/tree/main/examples/clients/simple-chatbot/mcp_simple_chatbot

[omri-madar]

Great, thanks!

[DawoodIrzam7]

Hi, I'm integrating the Atlassian MCP Server with a Node.js application. The MCP Server runs inside a Docker container, and currently we provide Jira credentials (such as JIRA_URL and JIRA_TOKEN) via environment variables at container startup.

Our use case requires supporting multiple users, each with their own Jira instance and credentials.

What’s the recommended approach to handle this kind of multi-user setup?

Is it possible to pass user-specific Jira credentials dynamically with each request to the MCP Server?

Or is it necessary to run a dedicated MCP container per user with their credentials set via environment variables?

[sooperset]

Hi, @DawoodIrzam7. We're planning to implement the multi-user support feature (relevant issue: #341).
Currently, the only way to do this is dedicated container per user as you mentioned!