At first glance, Qodo, the startup previously known as CodiumAI, may appear to be yet another AI code generation tool. But the team, which is announcing a $40 million Series A round led by Susa Ventures and Square Peg today, is just as much focused on generating code as on generating the tests and code quality in general.
The service launched out of stealth 18 months ago and so far, the company says, over 1 million developers have tried its solutions and a number of Fortune 100 companies have adopted its enterprise platform.
One major differentiator for Qodo is that while the company offers the usual extensions for Visual Studio Code and JetBrains to bring its Qodo Gen tools right into the IDE, it also offers a git agent that supports GitHub, GitLab, and Atlassian’s BitBucket, as well as a Chrome extension and a CLI tool.
“We call ourselves the first quality-first code generation platform for complex code,” Qodo CEO and co-founder Itamar Friedman told me. “In order to enable quality-first code generation, we believe we need to integrate into the entire software development lifecycle.”
Image Credits: Codium Friedman noted that each of the company’s tools aims to reduce bugs and issues in another part of the coding and deployment process. “It’s death by 1,000 cuts and we want to deal with each one of them in our vision. So we need to meet the bugs and the issues, either where they were created or where they can be caught. So we’re integrating into many locations.”
Another interesting aspect of Qodo is that it offers developers the option to write out the issue they are trying to fix and how they want to tackle it in natural language. This way, Qodo knows what you are trying to do when you start working directly with the code. Friedman called this “task-aware coding.” The agent doesn’t directly try to fix the issue, but now the code-completion tool is aware of what you are trying to do and can focus its suggestions on that.
After writing code, developers can then access Qodo’s test generation service within their IDE.
Later on, once the code has been pushed to GitHub, Qodo’s tools can help the reviewer quickly see what was changed and what the possible issues are. Friedman told me that the team is actually thinking about extending this tool to automatically create a video that walks the reviewer through the changes. “The human is still involved, but I need to make it easier for him or her to review [the code],” Friedman said.
Image Credits: Qodo By covering the entire lifecycle, Friedman argued, the individual tools can also learn from each other. If your reviewers always give you very similar comments about your code then the code suggestion tool can take that into account, for example.
“AI agents play an increasingly pivotal role in software creation, and we believe a quality-first approach is key for their widespread adoption at the enterprise,” said Yonatan Sela at Square Peg. “Devs at the enterprise don’t ‘start from scratch,’ their code needs to work in harmony with tens of thousands of lines of code that are already there.”
Looking ahead, the Qodo team plans to extend its service to also test the code from the user interface. This new service, tentatively called Qodo UX, will behave like a human who tries to use a company’s website and test the user interface for bugs.
“Someone could say, ‘Okay, just do end-to-end testing. But if you do end-to-end testing and you find a bug, you don’t know where it is, and you can’t really prove that you tested everything. Each testing type has its unique pros and cons. We want to cover all. We started with unit and functional [testing]. Then we grew to integrity. Next quarter is UX testing.”
In addition to Susa Ventures and Square Peg, Firestreak Ventures and ICON Continuity Fund also participated in this round, as did seed investors TLV Partners and Vine Ventures. This $40 million Series A round now brings the company’s total funding to $50 million.
CodiumAI is using generative AI to help developers build code logic tests automatically