Mihir Garimella
Actively Named a Launch Partner for OpenAI's New Agents SDK
Table of Contents
OpenAI released the next evolution of its Agents SDK today. The update adds native sandbox execution and a model-native harness for long-running agents. Actively was one of the teams that tested it before launch, and OpenAI featured us in the announcement alongside Oscar Health, LexisNexis, Thomson Reuters, and Zoom.
Long-running, reliable agents are the foundation of what we build. Every account gets its own agent that remembers what happened, notices what changed, and works out what should happen next. That only works if agents can run for long stretches, keep state, and work safely across many tools and systems.
"To run per-account agents at massive scale for revenue teams, we need flexibility, control, and room to customize. OpenAI's Agents SDK gives us that foundation, with an open-source design and first-class support for custom skills, hooks, and extensible memory. It lets us tailor agents to specific requirements—whether speed, cost, or intelligence—and scale with much more confidence and far less guesswork." Mihir Garimella, CEO, Actively
Read OpenAI's announcement: The next evolution of the Agents SDK. Read more about Per-Account Agents.


