AI tools have taken the world by storm. In April, we announced Atlassian Intelligence to bring the power of AI into our tools. Leveraging AI through internal models and our collaboration with OpenAI, Atlassian Intelligence will be built into the Atlassian suite of tools, including Bitbucket Cloud.
It’s still early days in understanding the full potential of AI, but we know that integrating AI into your development workflow can help you write code faster, automate repetitive tasks, and help you find and fix bugs before you ship. In this blog, we want to highlight apps by our AI partners that you can use today.
Here are three code-completion apps in the Atlassian Marketplace that integrate with popular IDEs and work across major coding languages. These tools can be integrated with your Bitbucket Cloud repositories to provide customized code suggestions to help you ship quality code faster. All of these apps also have on-prem versions that you can use with Bitbucket Data Center.
Codeium
Codeium is an AI-powered toolkit for developers. It can generate autocomplete suggestions, explain unfamiliar codebases, modify and refactor code based on prompts, and can translate code from one language to another. They also offer natural language-based search functionality, and a chat interface that works alongside your code so you can interact with it as you’re coding. You can personalize it by configuring it to personalize suggestions based on your Bitbucket repos, while your code is secure and remains within your tenant.
They support integrations with 40+ IDEs and 70+ languages. They had the widest range of support for IDEs and languages among the tools we tested. We tested it on Vim and found suggestions to be fast and accurate.
Codeium uses permissive open-source libraries to train their LLMs so you can be assured that code suggestions output by the tool is licensed for use.
Choose between cloud or on-prem versions.
Tabnine
Tabnine is a coding assistant that is trained on permissive open-source repositories – this means that the code suggestions provided by the tool are licensed to use, eliminating any future legal risks.
Tabnine provides code suggestions inline as you’re coding, including full-function code blocks, or you can generate code blocks based on natural language prompts. They use language-specific AI models, which help provide fast and accurate code suggestions. If you choose, you can also finetune suggestions based on your code in Bitbucket Cloud, without any IP exposure outside your systems.
We love that they support 30+ IDE’s so you can use almost any of the popular IDE’s. We tested it on VS Code and found their suggestions to be fast.
They offer a free trial of their Pro version for Bitbucket customers. They have cloud or on-prem versions.
Cody by Sourcegraph
Sourcegraph analyzes your codebase and creates a code graph, which is an indexed view of your codebase. Their new AI assistant Cody, uses a combination of your code graph, built from your Bitbucket repositories, combined with responses from AI models from OpenAI or Anthropic to make code suggestions. As they say on their site “You can think of Cody as your coding assistant who has read through all the code in open source, all the questions on StackOverflow, and your own entire codebase”.
Cody currently works on VSCode and IntelliJ with more IDE support on the way. We loved their Commands dropdown which has a great list of shortcuts (shown below) that you can use right away.
Integrating with GitHub Copilot
We know that many of our customers want all the advantages and benefits of being standardized on the Atlassian platform with Bitbucket, while still being able to take advantage of using GitHub Copilot.
In this blog, we’re excited to demonstrate how easy it is to use GitHub Copilot alongside your Bitbucket repos. Additionally, we’ll show how you can help keep developers in the flow by bringing context from across Atlassian tools right into your editor with the Atlassian Rovo for GitHub Copilot extension.
Atlassian Intelligence features in Bitbucket Cloud
Once code is pushed into Bitbucket Cloud, you can continue to use AI features within Bitbucket to help you accelerate PR cycle times and get to production faster.
- AI PR descriptions: Atlassian Intelligence analyzes your code changes and commit messages to craft a clear and concise pull request description so your reviewers have full context of code changes. Learn more
- AI code review: To accelerate code reviews, you can add the AI code review pipe into your CI/CD pipeline. With this pipe, AI will act as a code reviewer and leave comments in your code and provide suggestions for improvement based on syntax and code conventions. In the coming year, we are working on making AI code reviews part of the native experience.
- AI Chat (on the roadmap): We are also working on bringing you AI chat powered by Atlassian Rovo inside Bitbucket Cloud. With Rovo chat, you can use AI to get context from Jira or Confluence, right from within Bitbucket. For example, you can prompt AI with questions like ‘what Jira issues are linked to this PR’ or ‘summarize the code changes in this PR.’
Stay tuned for more AI updates!