Monitoring and learning layer for long-running agents
BentoLabs is the monitoring and learning layer for long-running agents. We detect when agents silently fail or drift from the user's goal, system prompt, or tool contracts, show affected users and root cause, and suggest the prompt, skill, or harness fix. As more teams deploy agents, keeping them reliable in production becomes mission-critical. Bento sits directly in the production loop and gives teams the operational leverage required to scale agent ecosystems without scaling human firefighting alongside them. The result is a system that turns opaque agents into agents that can be monitored, debugged, and improved continuously. The founders learned this problem at Emergent (YC S24), where they built and operated production coding agents used by 5M+ users. Abhinav was hire #1 and helped Emergent hit SWE-Bench #1 and scale from $0 to $100M ARR in just 8 months. Kaushik was hire #2, led full-stack engineering at Emergent, and was key to building the infrastructure that made production agents reliable, observable, and debuggable. Bento's self-learning engine has also lifted ARC-AGI-3 (internal) by 2.6x and Terminal-Bench 2.0 (internal) from 42.2% to 52.4% pass@1 with the same model, tools, and budget.
Founder
Kaushik was hire #2 at Emergent (YC S24), where he led full-stack engineering, built the initial infrastructure for long-running agents, and shipped the mobile-dev agent that converted ~50% better than any other. Previously Decathlon, Pazcare, and Synup.
Founder
Previously- Hire #1 at Emergent (YC S24). Abhinav led the Agents team, helping scale from $0 to $100M ARR in just 8 months and hit #1 on SWE-Bench, twice. Unofficially he was called ‘the agent whisperer’. He built BentoLabs after realizing that for Agents whatever he couldn't see it wouldn't get fixed. And the current monitoring tools were not delivering the value they promised. So he built the layer that actually finds the silent failure, fixes them and closes the loop.
