Raindrop, a monitoring platform designed specifically for AI agents, has raised $15 million in seed funding. The round was led by Lightspeed Venture Partners, with participation from Figma Ventures, Vercel Ventures, and YC. Notable angel investors include the founders of Replit (Amjad Masad and Michele Castata), Cognition (Walden Yan), Framer (Koen Bok and Joen van Dijk), Speak (Andrew Hsu), and Notion (Akshay Kothari).
The company positions itself as "Sentry, but for AI agents," referring to the popular error tracking tool used by software developers. Raindrop plans to use the funding to meet growing enterprise demand and establish itself as the go-to monitoring solution for production AI agents.
AI agents have become significantly more capable in recent months. They reason longer, use more tools, and connect to MCP servers. Some now run autonomously for hours at a time. This creates a problem: traditional testing methods simply cannot keep up.
"AI agents are more capable than ever. They're reasoning longer, using more tools, and connecting to MCP servers. Some agents now run autonomously for hours," said Raindrop CEO Zubin Koticha. "Traditional testing methods, like evals, aren't capable of handling the complexity of these long trajectories. This leaves engineering teams in the dark - unable to discover or track issues. Raindrop is the first monitoring platform to solve this problem."
The stakes are high. AI agents are increasingly deployed in sectors like healthcare and financial services, where failures can mean serious business consequences or even life-threatening situations. Recent incidents have made headlines: ChatGPT reportedly encouraged users to stop taking medication, and Air Canada faced a lawsuit after their chatbot promised refunds the company couldn't honor.
Raindrop will use the new capital to scale its operations and continue defining what production monitoring for AI agents should look like. The company already works with frontier AI customers processing millions of events daily.
Legacy monitoring tools typically offered basic metrics like latency and token usage. Some provided generic signals around toxicity or user sentiment. None of these effectively revealed actual errors in agent behavior. Raindrop takes a different approach: small, custom models that adapt to the unique characteristics of each AI product.
This allows engineering teams to monitor signals that were previously invisible. Beyond default signals like "User Frustration," customers can define custom alerts for specific issues such as "UI Aesthetic Complaints" or "Agent Stuck in a Loop." They can then track incident rates across millions of events. When problems emerge, the platform triggers notifications similar to Sentry's alerts.
The platform uses its own AI agents working in the background to triage and investigate potential issue patterns. When a new problem surfaces, such as an agent encouraging harmful user behavior, Raindrop collects relevant information and generates step-by-step explanations of what happened.
Many teams also use Raindrop Experiments, the company's A/B testing platform for agents. This lets them validate solutions, whether tool changes, model swaps, or broader pipeline modifications, and confirm their fixes actually worked.
"Raindrop has been invaluable as we've been growing quickly," said Evan Goldschmidt, CTO at Tolan. "It's critical for us to keep issue incidence below an acceptable threshold and become aware of any spikes. It's like if we see an iOS crash report in Sentry, but for our AI capabilities."
Raindrop was founded by Ben Hylak, Zubin Koticha, and Alexis Gauba. The founding story is practical: the team was building their own coding agent and kept running into silent failures in production. During Y Combinator, they noticed everyone building agents had the same problem.
Koticha and Gauba are second-time founders whose previous company was acquired by Coinbase. Hylak spent time as an engineer at Apple before joining Apple's Human Interface Design team, where he worked for four years.
Lightspeed Venture Partners led the round. The firm, which manages $30 billion in assets, has backed companies including Anthropic, Snap, Affirm, Epic Games, Mistral, and Wiz over its 25-year history.
"We keep seeing AI engineering teams struggling with agent failures in production, and traditional evals are not really helping," said Bucky Moore, Partner at Lightspeed. "We invested in Raindrop because they defined monitoring for AI agents. They made an early bet that monitoring would be the most critical part of building reliable agents, and they've been right. They're growing fast as Raindrop is being embraced by some of the most forward-thinking engineering teams in AI. Lightspeed is thrilled to partner with Alexis, Ben, Zubin, and team to help them scale Raindrop into the default agent observability platform for enterprises and startups alike."
The round's investor roster reflects strong interest from the AI and developer tools ecosystem. Figma Ventures and Vercel Ventures bring expertise in design and frontend infrastructure, while the angel investors from Replit, Cognition, Framer, Speak, and Notion represent companies actively building AI-powered products themselves.
