This blog was written by Jeff Wang, Head of Business at Codeium. Please tag @Jeff Wang - Codeium when leaving comments. We frequently review the latest code AI apps and found Codeium to have widest range of support for IDEs and languages among the tools we tested. Our developers found the suggestions and chat to be fast and accurate. Codeium can also directly connect to your Bitbucket repositories for better context awareness for both autocomplete and chat. |
AI coding assistants are more popular than ever, with some companies claiming that more than half their new code is being generated by coding assistants.
With recent announcements of GitHub Copilot integrating features such as code base personalization and GPT4 chat, Bitbucket users may feel like they are missing out on this wave of AI. That’s not true. Coding assistants work in the IDE, so any Bitbucket customer can integrate an AI coding assistant into their IDE and connect to their repos for better contextual assistance.
In this blog, we’ll walk you through how to use Codeium’s AI tool with your Bitbucket repos. Codeium integrates with all the popular IDEs.
Codeium is one of the fastest moving AI startups in the coding assistants space. Having reached 300k monthly active users, and with major enterprise customers such as Dell, Anduril, and others, it’s a trusted AI platform. Codeium is an Atlassian partner and you can get the app in the Atlassian marketplace.
Codeium takes IP security very seriously. We trains AI models on code that can be used without any copyright restrictions. And when we connect to your repos, we don’t retain any data. If you’re using our self-managed tool, all your code remains fully air-gapped in your environment.
In this blog, we’ll cover how Codeium works with Bitbucket to deliver an end to end AI assisted experience.
Let’s say you open up the Jira issue assigned to you, and you have no idea where the code is.
From this Jira issue, let’s checkout the branch from Bitbucket:
I’ve opened up my VS Code and have switched to this new branch, but have no idea where this file is! As you can see I am greeted by Codeium Chat inside my IDE:
Because Codeium is integrated into your IDE, it can index your Bitbucket code locally for any chat or auto-completions. You can even connect your Bitbucket repo to Codeium Enterprise for fine-tuning via read-access tokens. For this example, I’ve asked my codebase on Bitbucket directly, “where is my contact form located?”
It has returned the correct result and a few other suggestions:
Amazing! Now I can simply just start typing what I want:
All I need to do is hit “Tab” to accept suggestions, or cycle through multiple selections with Option/Alt + [
After only writing a comment, I was able to accept the rest of the suggestions. This is why hear of metrics such as coding assistants are writing more than half of the code being produced! Better yet, I did not have to reference other websites to look up Regex at all.
Codeium has context awareness across your entire Bitbucket repo. I can simply type an “@” symbol in the chat to chat about my Functions, Classes, or even other Repos!
From this screenshot, you can see that the chat was able to give me a solution when asked about precompiling the regex in the code that was accepted.
Finally, I can round out my code by generating unit tests or even docstrings with the click of a button or using chat.
To summarize, Codeium provides context awareness and search, along with many advanced chat features that already work well with Bitbucket today. Not only that, we were able to use chat to further bolster our code with Unit Tests and Docstrings.
To use Codeium with Bitbucket on your favorite IDEs, find us at on the Atlassian marketplace.
Ash Moosa
3 comments