Blog/

case-studies

StarSling: Building Cursor for DevOps with Mastra

·

Jul 1, 2025

Daniel Worku was supporting 3,000 engineers at Netflix when he spotted a problem.

He'd built Netflix Console — the tool that helped ship ads, live streaming, and games in record time. But engineers still wasted 20-30% of their time on boring tasks outside their code editor. Bug fixes. Deployments. Performance issues. Production fires.

"We had this tool for tracking services and pinging engineers, but it had no AI," Worku recalls. "It was basically a manager's bookkeeping tool."

That's when Worku called his old friend Yonas Beshawred.

Beshawred had just sold StackShare, his developer tools company, to FOSSA. The two had met at a Bay Area hacker house ten years earlier, bonding over Ethiopian culture at a party. Beshawred was even a groomsman at Worku's wedding.

"When we talked about building an AI-powered tool for engineers, Yonas was like, 'I'm down to do it again,'" says Worku.

Now part of Y Combinator's S25 batch, StarSling turned to Mastra in early 2025. Their goal: bring Cursor's AI magic to all the work that happens outside the code editor.

Yonas Beshawred and Daniel Worku, co-founders of StarSling: from groomsmen to co-founders

Yonas Beshawred and Daniel Worku, co-founders of StarSling: from groomsmen to co-founders.

The Vision: Cursor for Everything Outside the IDE

StarSling targets the 20-30% of engineering work that happens after you write code. "Cursor transformed how engineers write code," Worku explains. "But step outside the IDE and it's like going back to the stone age."

"You finish coding. Now you need to deploy. Or fix a slow database query. Or debug a production error. That's what we're building for—all those annoying alerts that interrupt your day."

The team is starting simple: a prioritized inbox for GitHub, Linear, and Sentry. See an exception? Click "Autofix." An AI agent debugs it, finds the root cause, and opens a pull request. All from your homepage.

Starsling Autofix feature in action

StarSling's Autofix feature in action.

Why Mastra Powers StarSling's Agent System

StarSling needed to build multiple AI agents. Each would handle different tasks but work together across developer tools.

"We evaluated LangGraph and considered building something custom," says Worku. "But Mastra's modular agent framework and workflow engine were perfectly suited for what we needed."

Here's how StarSling uses Mastra:

Frontend: Like Vetnio, StarSling uses assistant-ui, a responsive frontend with deep Mastra integration that works seamlessly on desktop and mobile.

Agent Orchestration: Mastra's workflow orchestrator routes tasks to specialized agents based on the type of task — whether it's fixing a Sentry exception, updating a Linear task, or opening a GitHub PR.

Cross-Tool Integration: Mastra enables agents to work across multiple developer tools. "Fixing exceptions potentially requires getting data from Sentry, opening a GitHub PR, and linking to a Linear task" Worku explains. "Mastra makes this orchestration seamless."

Infrastructure: StarSling runs serverless on Vercel and Cloudflare. They're also big believers in open source — alongside Mastra, they use Better Auth for authentication and Inngest for data syncing.

Looking ahead

Mastra's docs and Slack support helped StarSling ship fast.

"If you have questions, beyond getting a fast response from Abhi or Shane, you can look at the code yourself," notes Worku. "Open source is a huge unlock."

Instead of building infrastructure, they focused on their product. The autofix feature is just the start—they plan to automate all DevOps tasks. The team is rapidly growing; StarSling is looking to hire their first engineers.

"Every engineer who has code in production does this type of work," Worku emphasizes. "Especially when you're on-call, this 20-30% of your time becomes 60-70%. We're building Cursor for DevOps—bringing that same AI-powered experience to everything that happens after the code is written."

Stay up to date