OpenAI and Cursor agree on agent coordinators. They disagree on who runs them.
OpenAI opened its Agents API in public beta this month, exposing the harness that powers Codex with managed sessions, tool The post OpenAI and Cursor agree on agent coordinators. They disagree on who runs them. appeared first on The New Stack .
OpenAI and Cursor have unveiled coordinated coding agents that orchestrate multiple specialized agents to tackle larger software development tasks. While both products share a similar architecture, they differ in their approach to who controls the agents. OpenAI's Agents API exposes a harness for Codex with managed sessions, tool coordination, and subagent orchestration, allowing the coordinator to enforce guardrails and route work to the most efficient agent.
Cursor's Projects feature similarly orchestrates multiple coding agents to work together on larger bodies of software work. Both solutions address the issue of maintaining reliability in multi-agent workflows, as a single agent can struggle to maintain relevant information across numerous stages of a complex software development task.
The emergence of these coordinator-worker models highlights the need for reliable systems when working with multiple capable agents on the same problem.
Written by urgent.news from The New Stack's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
