What's New in CrewAI: OSS Updates & the Agent Management Platform

Join Jesse Miller (VP of Product) and Lorenze Hernandez (Lead OSS Engineer) for a technical walkthrough of everything shipping across the CrewAI stack — from the open source framework to the enterprise platform.

Lorenze will cover the major additions from the 1.14.x line, with runtime state checkpointing and fork-and-resume support as the headliner. He'll also walk through expanded async support across crews, flows, tasks, and tools, first-class MCP and A2A integrations, and new sandbox tools for E2B, Daytona, and Bedrock V4.

Jesse will take you through the Agent Management Platform (AMP) — the enterprise layer for shipping agents to production. He'll focus on the first three pillars of AMP — Discover, Build, and Deploy — and how they help teams find the right automations, ship them faster, and run them in production. Plus a look at the new Skills Repository for sharing and reusing agent capabilities across your org.

What you'll get:

• A working understanding of checkpointing and fork-and-resume — and how to start using it today

• A tour of what's new in async, MCP, A2A, and sandbox tooling across the 1.14.x line

• A walkthrough of AMP's first three pillars: Discover, Build, and Deploy

• A first look at the Skills Repository and how it changes agent reuse at scale

• A look at how Discovery surfaces where to automate before you write a line of code