Artificial intelligence is transforming the coding world, offering powerful tools that redefine how developers troubleshoot and optimize code. From GitHub’s Copilot Workspace and Amazon’s CodeWhisperer to Oracle’s Code Assist, Anysphere, and Cognition’s Devin, there’s no shortage of ways to churn out code to accelerate product development readily. But that’s only one part of the puzzle—teams must still review the code to ensure it does what was intended and is error-free. And that takes time. A startup called CodeRabbit believes its solution can expedite things, automating the code review process and ensuring high code quality.
This week, the company announced a $16 million funding round to support its mission of having AI tackle code reviews. Venture firm CRV led the round, which included participation from Flex Capital, Engineering Capital, and angel investor Datadog Chief Executive Olivier Pomel, among others. CodeRabbit says it’ll use the new cash infusion to expand its product offerings and third-party integrations such as Jira, Slack, integrated developer environments (IDEs) and AI-powered analytic tools.
Making Code Reviews Faster
“Today, every company is a software company and every software developer writes code that must be reviewed,” Harjot Gill, CodeRabbit’s chief executive, told me in an email. “Code review is a critical component of the software development lifecycle but often fraught with inefficiencies and inconsistencies. Manual code reviews are time-consuming, prone to human error, and can create bottlenecks in the development process. As teams scale, these challenges multiply, leading to delayed releases, overlooked bugs, and inconsistent code quality.”
He adds that developers spend an average of nearly a third of their time on code reviews, which can take days to complete and bog down product delivery.
With CodeRabbit, companies can use AI to automate the process, freeing up developers to “focus on more complex and creative tasks, leading to faster delivery times, higher code quality, and a smoother development process overall.”
Integrated into existing workflows, developers can generate pull requests to merge their code into central repositories. CodeRabbit’s AI will review the changes and produce a summary with actionable feedback. Developers can accept these changes with a single click without jumping back into an editor to modify their code. For added support, a chatbot is included to allow developers to ask questions about feedback that might not meet their company’s coding practice.
The company boasts that its platform uses advanced AI reasoning, which can understand the intent behind the code and deliver “actionable, ‘human-like’ feedback.” They say this contrasts with current methods, including traditional static analysis tools and linters, which are “rule-based and often generate high false-positive rates, and peer reviews that are “time-consuming and subjective.”
Gill explains that CodeRabbit complements AI offerings from GitHub and others: “While tools like Copilot assist developers in writing code faster, CodeRabbit ensures that the generated code is thoroughly vetted for bugs, security vulnerabilities, and adherence to best practices before it’s merged.”
Sounds great, right? But what about the risk of hallucinations? If developers are using AI assistants to generate their code and conduct reviews—even though they’re different AIs—are we sure that the output will be free of vulnerabilities? Who’s watching the watchers?
CodeRabbit to the Enterprise
Since being co-founded by Gill, a former Nutanix executive, and Gur Singh, a former Alegeus executive, CodeRabbit has signed up more than 1,000 organizations, most of which are small—to medium-sized businesses. It has recently been making inroads into the enterprise, onboarding “several Fortune 500 companies” and planninge its marketing efforts.
In early proof-of-concept trials, the startup learned these enterprise firms hesitated to embrace its Software-as-a-Service (SaaS) product and transmit their data to the cloud. It might have been because these organizations were particular about having their data in a location they could control. To overcome this, CodeRabbit launched a self-hosted solution, allowing its large customers to run within their infrastructure.
With the software development lifecycle notoriously lengthy in the enterprise, one could assume that deploying AI to accelerate code review could boost innovation and help organizations ship products faster. It’s possible that companies could program CodeRabbit to be cognizant of all their coding, legal, and security practices so the AI will be able to effectively and efficiently parse through submissions and reduce production time.
But while it’s touting its AI efforts as revolutionary, CodeRabbit isn’t alone in its mission. Competitors include SonarSource, Codacy, Snyk, Codium AI, Ellipsis, and Greptile AI. However, CodeRabbit believes it’s “significantly ahead in terms of traction.”
“Code review has already been a big pain point in the developer workflows and it is just going to get worse with large amounts of code getting written with the AI completion tools like GitHub Copilot,” Gill remarked. “AI will augment the large part of the manual review process.”
He added, “The speed of software has often been a hurdle, but we believe AI will change that…AI has proven its potential in developer tools. We see AI playing a role in every step of the developer workflow in [the] future, from code generation to code reviews and testing and validation.”
Leave a Reply
You must be logged in to post a comment.