Lightrun Launches Its AI Debugger to Help Developers Fix Production Code

LightrunA Tel Aviv-based startup that helps developers debug production code from inside their IDE announced Wednesday the launch of its first AI-powered tool: Runtime Autonomous AI Debugger. The recent tool, currently in private beta, goals to help developers fix issues with production code in minutes, not hours.

Additionally, Lightrun also disclosed on Wednesday the $18 million SAFE it raised last yr from GTM Capital, with participation from existing investors Insight Partners and Glilot Capital. Lightrun’s total funding to date is $45 million. The company is reportedly planning to raise a Series B round next yr.

“We used to get (average recovery time) down to, let’s say, 30 minutes, maybe 45 minutes on average, based on how we evaluate ourselves and what our customers say,” Ilan Peleg, CEO and cofounder of Lightrun, told me. “Now we’re going to automate everything from the moment a ticket comes in all the way to finding the root cause down to a single level of detail, like which of your individual lines of code is responsible for that particular root cause.”

In time, Peleg said, Lightrun would love to expand this to include using generative AI to routinely fix bugs. That’s impossible yet, but given how quickly the technology has advanced, it’s likely only a matter of time.

To do that, Lightrun is tuning existing models to deal with debugging, which the corporate can do partly since it gets information not only from the code itself but from its entire monitoring and commentary stack. Looking ahead, the corporate also plans to connect the system to other enterprise inputs, comparable to ticketing systems. “There’s so much data in the enterprise environment that is somehow related to troubleshooting or debugging—and that’s missing from solutions like Copilot,” Peleg said. Most chat interfaces like Copilot, he argued, just take a look at the code but don’t have enough context to provide one of the best solutions.

As Peleg notes, the team went through several iterations before it felt its system was ready for on a regular basis use. About six months ago, Lightrun began experimenting with existing models to see where generative AI could help its users. At the time, nevertheless, the answer was far too expensive to offer as a product. “We’ve now fine-tuned our system … so that it doesn’t have a significant cost to us as a solution, which is why we’re talking now. In the past, I’ve been uncomfortable announcing something that wasn’t there yet.”

For now, at the least, these generative AI features will simply be a part of Lightrun’s existing experience for users in private beta. Peleg emphasized that the corporate wants to prove that the system actually delivers value to users and isn’t trying to optimize it for short-term monetization.