Integral, a venture-backed data management startup, today debuted an open-source tool that can analyze software teams’ code and automatically highlight areas for improvement.

The tool is known as Robin AI. It uses artificial intelligence to scan code files for potential issues.

Integral, officially Promeed Inc., launched in 2022 and earlier this year raised a funding round of undisclosed value. It’s backed by several institutional investors including LiveRamp Holdings Inc., a publicly traded provider of data management software.

Integral provides a platform that helps healthcare organizations ensure their data analytics initiatives comply with regulations. The startup built Robin AI, the open-source tool it debuted today, to make its software engineers more productive. Integral claims the tool has helped reduce the number of bugs that its developers must fix manually. 

“Because it serves as a trusty partner in enhancing the code development processes, we named the project after Batman’s sidekick Robin,” said Integral co-founder and Chief Technology Officer John Kuhn. “It has served us well in boosting product velocity and reducing production bugs for Integral’s engineering team.”

Robin AI activates when developers submit a so-called pull request. That’s a notification indicating a newly created code file is ready to be deployed, but must first be double-checked for potential issues. It uses a machine learning model from OpenAI LP’s GPT series of neural networks to automatically scan code files for bugs after pull requests. 

Certain snippets of code are expected to work reliably even if malfunctions occur. To achieve the necessary degree of reliability, developers must implement so-called error handling mechanisms in their software. According to Integral, Robin AI can detect if a code file is missing error handling mechanisms and also spots a range of other common software faults. 

The tool is capable of identifying more subtle issues as well. It can, for example, spot code snippets that don’t have sufficiently descriptive names. Giving a code construct a name that explains its purpose makes it easier for developers to understand.

After identifying potential issues in a code file, Robin AI creates a list of improvement suggestions. It also provides code samples that illustrate how those suggestions can be implemented. For added measure, it generates a code quality score ranging from 0 to 100 that indicates the urgency of the detected issues. 

Robin AI is implemented as a Docker container, which makes it relatively simple to deploy. It’s designed to work with GitHub’s GitHub Action tool. The tool enables developers to run automated software workflows that interact with their code repositories in various ways, such as by highlighting optimization opportunities. 

Robin AI has released the source code on GitHub. The tool is available under a free MIT license. 

Photo: Unsplash