Developers use Bitbucket Cloud every day to turn plans into working software. Code lives in repositories, feedback happens in pull requests, and Bitbucket Pipelines helps teams validate changes before they move forward. For an AI assistant to help with this work, it needs access to the same information. Without a direct connection, developers have to gather pull request details, pipeline results, logs, and links before the assistant can help.
Bitbucket Cloud was already available through the Atlassian MCP Server using API tokens. In this latest release of Atlassian MCP, developers can also connect through OAuth by signing in with their Atlassian account. API tokens remain supported, while OAuth offers a simpler way to get started with compatible AI clients.
Bring work and development context together
Once connected, an AI assistant can access both the work context behind a change and the development context needed to implement it. Jira explains why the change is needed, Confluence captures technical decisions and rollout plans, and Bitbucket provides the code, commits, pull request discussion, and pipeline results.
With all this information available through one connection, developers can start with the task instead of copying details between tools.
Atlassian MCP also expands how an assistant can work with Bitbucket. It can manage pull request comments and reviewer updates, find commits, and locate related pull requests. This lets the assistant help with implementation and review, not just retrieve information.
Here is how that can work for two common development tasks.
From Jira and Confluence to an implementation in Bitbucket
The reason for a code change often lives outside the repository. A Jira work item may describe the customer problem and acceptance criteria, while a linked Confluence page explains the technical design or rollout plan.
With Atlassian MCP, a developer can bring that information into their AI coding workflow and use it to guide the implementation in Bitbucket. The assistant can identify the areas of code likely to be affected, help implement the change, and prepare a pull request for review.
From a failed pipeline to a reviewable fix
When a Bitbucket Pipeline fails, the first challenge is understanding what went wrong. A developer may need to check the failed step, read its logs, review recent changes, and inspect the affected code.
With Bitbucket connected through Atlassian MCP, an AI assistant can bring those details together, suggest a likely cause, and help prepare a focused fix. The developer can then focus on validating the diagnosis and reviewing the proposed change.
Work from the AI client you prefer
Development teams use different AI clients depending on the task and how they prefer to work. Atlassian MCP makes Bitbucket information available in compatible clients, so developers are not limited to a single interface.
Whether you’re using an IDE such as Cursor or VS Code, or an AI client such as Claude or Codex, you can work with repository, pull request, and pipeline information in the tool you already use. Access remains governed by your existing Atlassian and Bitbucket permissions.
Start using Bitbucket with Atlassian MCP
Bitbucket Cloud OAuth support is available through the Atlassian MCP v2 endpoint: https://mcp.atlassian.com/v2/mcp
To get started:
- Confirm that your Bitbucket Cloud workspace is linked to an Atlassian organization.
- Add the Atlassian MCP Server through your preferred AI client’s plugin, connector, or server configuration.
- Sign in with your Atlassian account.
- Review and approve the requested access.
- Try one of the workflows described above.
Visit the Atlassian MCP getting-started guide for setup instructions for supported clients.
If you currently use Atlassian MCP v1, follow the v2 upgrade guide. Depending on your client, you may need to remove and add the existing plugin or connector again, then sign in with your Atlassian account.
Help shape what comes next for Bitbucket
OAuth makes it easier to connect Bitbucket Cloud to AI assistants. As AI continues to part of development work, we expect more developers to interact with Bitbucket through agents and workflows. To support this shift, we’re making more Bitbucket capabilities available through APIs, Atlassian MCP, and TWG CLI. This includes understanding what is blocking a pull request, working with merge queues, and reviewing deployment and test history.
We want to understand which workflows matter most to your team. Which Bitbucket tasks would you like an AI assistant to help with? Share your ideas and feedback in the Bitbucket Cloud Community.
