Laminar, a startup focused on observability and debugging for AI agents, has announced a $3 million seed funding round led by Atlantic.vc, with participation from Y-Combinator and other investors.

Observability for AI Agents

The company aims to solve the challenges related to the observability of AI agents, which often perform complex tasks for extended periods, generating large amounts of data. Traditional observability tools are not designed to handle this complexity, making it difficult to identify and fix errors.

Laminar offers a solution that captures every action taken by the agent, including calls to large language models (LLMs), the use of external tools, and the execution of functions. For agents operating in browsers, Laminar also records sessions, synchronizing them with execution traces to provide a complete visual context.

Key Features

  • Signals: Uses AI to automatically identify error patterns and anomalies.
  • Agent Debugger: Allows developers to rerun an agent from any step, preserving the previous context.

Laminar's customers include Browser Use, OpenHands, and Rye.com. Laminar's SDK is integrated into OpenHands' software agent and Browser Use's documentation.

Future Prospects

The funding will be used to accelerate product development and go-to-market expansion, at a time when AI agents are moving from the prototype phase to large-scale implementation.