Coordination, not a single agent

AI agent orchestration, explained.

Orchestration is what turns a single assistant into a team: coordinating multiple agents, sequencing steps, and passing context between them. Here's what it means and which open-source frameworks do it.

Grounded in public docs · September 2026

The concept

What orchestration is

A single agent does one thing at a time. Orchestration coordinates several agents toward one outcome.

  1. Role assignment. A researcher agent, a writer agent, a reviewer agent — each with a defined job.
  2. Sequencing and handoffs. Output of one step becomes the input to the next.
  3. Shared context. Agents pass state and memory so the team stays coherent.

Open-source frameworks

The frameworks that do it

MIT

CrewAI

"Framework for orchestrating role-playing, autonomous AI agents." A Python library for building teams of agents that collaborate — ~59k stars.

Microsoft

AutoGen

"A programming framework for agentic AI." Multi-agent conversations for complex workflows — ~61k stars.

agent runtime

OpenCode

The open-source agent CLI that Kortix and OpenWork build on. Agents with a scoped reach into tools, on isolated sandboxes.

Framework vs platform: CrewAI/AutoGen give you a library and leave hosting, memory, secrets, and review to you. A management platform (Kortix) bundles orchestration with a review gate and audit trail.

When you need it

Single agent vs multi-agent

One agent is enough for a single bounded task. Orchestration earns its complexity when the work is naturally several roles — research, synthesis, and review — that benefit from specialization and handoffs. Start simple; add orchestration only when a single agent's output quality plateaus.

Compare platforms and frameworks →