Swiss startup LogicStar is stepping into the AI game with a fresh twist. Founded in summer 2024, the company has just raised $3 million in pre-seed funding to create tools that handle the autonomous maintenance of software applications. Unlike typical AI agents that focus on code creation, LogicStar is targeting a different pain point: fixing bugs and keeping apps running smoothly.

The idea is simple: let AI handle the tedious stuff so developers can focus on the fun part—building new features. Think of it as a tireless assistant that spots and solves problems in deployed code. And the potential is huge. Experts say even the best AI models out there struggle to fix most bugs, which is where LogicStar sees its opportunity.

The company is building its platform on top of large language models (LLMs) like OpenAI’s GPT and China’s DeepSeek. By taking a model-agnostic approach, LogicStar can pick and choose the best model for each specific issue, maximizing efficiency.

Here’s how it works: the platform analyzes an application using “classical computer science methods” to create a detailed map of its structure. When a bug pops up, the AI agent narrows down the affected parts and tests thousands of potential fixes to find the best solution. It’s like having a super-speedy debugger that knows the app inside out.

LogicStar’s initial focus is on enterprises, offering its “silicon agents” to work alongside corporate dev teams. The goal? To save time and money while freeing up human developers for more creative tasks.

The startup is currently testing an alpha version with a few undisclosed companies and plans to expand its support beyond Python to include Typescript, Javascript, and Java soon.

With European VC firm Northzone leading the pre-seed round, LogicStar is well-positioned to make waves in the AI-driven software maintenance space. A beta release is expected later this year, and developers can already sign up for early access.

The future of software maintenance is looking smarter—and a lot less frustrating.