The Tool Desk
Outbyte Driver Updater FREEScan for outdated or missing drivers - takes under a minuteDriver Scan →Outbyte PC Repair FREEClear out junk files and repair common Windows errorsFree Scan →To connect an MCP-compatible AI client to Atlassian Cloud, add Atlassian’s hosted Rovo MCP server at https://mcp.atlassian.com/v2/mcp, then complete the client’s Atlassian OAuth 2.1 sign-in and consent flow. For an interactive user session, OAuth is the recommended route. API-token authentication is an optional, administrator-controlled choice for non-interactive automation.
This connection does not install a server in your environment: the MCP endpoint is hosted by Atlassian. The client connects to it remotely, and the tools it can use are governed by the signed-in user’s Atlassian access and your organization’s policies. Atlassian’s getting-started guide is the primary setup reference.
What you need before connecting
Use an MCP-compatible client that supports connecting to a remote MCP server and completing an Atlassian authentication flow. Atlassian documents setup routes for clients including VS Code with GitHub Copilot, Cursor, Claude Code, Claude Desktop, Codex Desktop and Windsurf; the exact screens and configuration fields depend on the client and may change as its MCP support evolves. When the client provides an Atlassian-specific installation flow, prefer that over manual configuration.
- An Atlassian Cloud account with access to the relevant site and products.
- Permission from your organization’s settings to use the MCP server and the selected external AI client.
- If you are connecting from a restricted corporate network, a permitted network or VPN address.
Atlassian’s remote server is the hosted endpoint; you do not need to deploy a separate MCP server just to use this connection. Consult the official Atlassian MCP Server repository for Atlassian’s client and administration context.
#1 Best Overall
Connect a client with interactive OAuth
- Open the client’s MCP or integrations setup. Choose its Atlassian/Rovo setup option if available. Otherwise, add a remote MCP server and enter
https://mcp.atlassian.com/v2/mcpas the server URL. Do not substitute an older v1 address when setting up a new connection. - Start the Atlassian authentication flow. In the client, select the option to connect or authenticate with Atlassian. The client should take you to Atlassian’s authorization and consent screen.
- Sign in and review consent. Sign in with the Atlassian account whose data the client should access. Review and approve the requested access only if it matches your intended use and your organization permits it.
- Return to the client and verify the connection. Finish any confirmation step shown by the client, then check that it reports the server as connected or makes Atlassian tools available. Client labels differ, so do not assume that a particular button name or configuration-file format applies to every client.
OAuth 2.1 is Atlassian’s recommended interactive authentication approach. Its OAuth 2.1 configuration guide describes the authentication flow. Authentication establishes the user identity; it does not override the permissions that user already has in Atlassian.
Manual endpoint or native setup?
Use the native Atlassian route when the client offers one, because it can guide the authentication handoff for that client. Manual setup is appropriate when the client instead asks you to provide a remote server URL. In either case, the remote server address is https://mcp.atlassian.com/v2/mcp; the OAuth interaction is completed through the client rather than by pasting a personal password into a generic URL field.
Rank #2
When to use the complete tool-list variant
Most clients can discover tools dynamically. If you are connecting through an MCP gateway that requires a complete, paginated tool list rather than dynamic discovery, Atlassian documents this endpoint variant: https://mcp.atlassian.com/v2/mcp?tools=all. Use it only when that gateway’s discovery behavior calls for it; it is not a different authentication method. See Atlassian’s endpoint and setup guidance.
Choose authentication for the way the connection runs
| Use case | Recommended approach | What to check |
|---|---|---|
| Person using an AI client interactively | OAuth 2.1 sign-in through the client | Confirm the client is using the Atlassian consent flow and the intended user account. |
| Backend, CI/CD pipeline, bot or other non-interactive process | API-token authentication only if the organization administrator has enabled it | Ask the administrator which credential type is permitted and how it should be provisioned and stored. |
Atlassian documents a personal API token sent with Basic authentication and a service-account API key sent as a Bearer token. These are not interchangeable descriptions of one credential: use the method associated with the credential issued for your scenario. Do not embed a personal token in client-side code, source control, logs or a broadly shared configuration. Store machine credentials in an approved secret store, restrict who can retrieve them, and rotate or revoke them according to your organization’s policy. Atlassian’s authentication and authorization guide explains the permission model, and its API-token setup guide covers the documented token route.
Do not select API-token authentication merely because a client offers a field for a token. Atlassian describes it as an optional route for non-interactive use that depends on administrator enablement. If the organization has not enabled it, use the supported interactive OAuth route or ask the administrator about the approved alternative.
Understand permissions, administration and safety
The connected client acts with the authenticated user’s existing Atlassian permissions. Connecting MCP does not itself grant broader access to Jira, Confluence or other Atlassian data. A user who cannot access a resource in Atlassian should not expect the MCP connection to make it available.
Rank #4
Organization and site administrators can manage or revoke the MCP app’s access and control which external AI tools or domains are allowed. Network/IP allowlisting can also prevent a valid client from reaching the service. If setup fails only on a corporate network, ask an administrator to check the currently permitted network or VPN address rather than repeatedly changing credentials. Atlassian describes these controls in its server repository and external MCP server administration guidance.
- Connect only a client your organization trusts; a connected AI system can invoke tools on the user’s behalf.
- Use an account with only the Atlassian access required for the task, especially for automated or broadly shared workflows.
- Review proposed high-impact changes before approving them, and monitor audit logs where your organization’s controls provide them.
- Consider prompt-injection and tool-poisoning risks: content encountered by an AI system may try to influence its actions, so do not treat retrieved content as inherently trustworthy instructions.
Atlassian explicitly cautions about the risks of connected AI clients and recommends taking care with access and actions. See its Rovo MCP guidance and security overview.
Quick wins for a faster PC:
Repair Windows errors before they cause bigger problemsFix Now →Fix the driver behind crashes, sound loss and screen glitchesFind Drivers →Best Value
Account for Rovo credit use
Some enriched Teamwork Graph, unified search and context calls consume Rovo credits. Consumption depends on factors including request complexity and how much context is fetched; the applicable allowance and thresholds depend on the Atlassian plan. There is no single universal allowance to apply to every account. Check the current plan and usage information for your organization before estimating the cost of a workflow that relies heavily on enriched calls. Atlassian discusses this in its remote MCP setup article and getting-started guide.
Troubleshoot connection and authentication failures
| Symptom | Likely check | What to do |
|---|---|---|
| The server does not connect or tools are not discovered | Endpoint or client discovery configuration | Confirm the URL is https://mcp.atlassian.com/v2/mcp. If a gateway requires the complete paginated list, configure the documented ?tools=all variant. Check the client’s own remote-MCP instructions for its expected discovery mode. |
| OAuth fails after moving from an older setup | Stale v1 configuration, cached client ID or cached .well-known credentials |
Reconfigure the client for the v2 endpoint. If authentication still fails, clear the client’s cached credentials or stale connection entry and start the Atlassian sign-in flow again. |
| The URL and account appear correct, but the connection is blocked | Organization policy or network/IP allowlisting | Ask the Atlassian organization or site administrator whether the MCP app and external AI client/domain are allowed, and whether your current network or VPN address is permitted. |
| “Invalid token” or “invalid context” | Token setup, authentication context or Rovo MCP Server settings | For API-token use, verify with the administrator that token authentication is enabled and that the correct credential type and method are configured. For persistent errors, have an administrator inspect the Rovo MCP Server settings and follow Atlassian’s invalid-token and invalid-context troubleshooting steps. |
| A tool cannot access a particular project, page or record | The signed-in user’s existing Atlassian permissions | Check access in Atlassian itself and confirm the connection authenticated the intended account. MCP does not independently expand that account’s permissions. |
When reporting an unresolved issue, give your administrator the endpoint, client name, whether you used OAuth or an enabled API-token method, the failure text, and whether the problem changes on an approved network. Do not send passwords, API keys or bearer tokens in a support ticket. Atlassian’s setup guide and troubleshooting article are the relevant official references.
Or skip the browser setup
ScreenshotNeo is a separate website screenshot API and MCP server, not an Atlassian connector and not a substitute for the Atlassian OAuth flow above. If your adjacent task is capturing a website page as an image or PDF, rather than querying Atlassian through MCP, one GET request can return a screenshot. For example, this captures Atlassian’s Rovo MCP getting-started page as WebP:
Quick Recap
curl -G "https://api.screenshotneo.com/v1/shot" -d access_key=YOUR_API_KEY --data-urlencode url=https://developer.atlassian.com/cloud/rovo-mcp/guides/getting-started/ -o shot.webp
See the ScreenshotNeo API documentation for the request options. ScreenshotNeo accepts cookie and consent banners like a visitor, then removes more than 60 known consent platforms, newsletter popups and chat widgets before capture; those steps can be turned off. Bot checks/CAPTCHAs, blank pages, timeouts, failed loads and cache hits are not billed, and responses identify the page verdict and billing status. Its MCP server offers take_screenshot, get_page_info and capture_pdf for Claude, Cursor and other MCP clients. The Free plan includes 1,000 screenshots per month with no card; paid plans start at $5 for 3,000, and every feature is available on every plan. Learn about ScreenshotNeo or sign up free for 1,000 screenshots a month with no card.
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.




