Playnex Docs – Orchestration

1. What Is Orchestration?

Orchestration is the coordination of multiple agents to achieve a larger goal. Each agent does what it’s best at — research, writing, planning, analysis — and Playnex manages the flow between them.

Instead of one overworked generalist, you get a small team of specialists, each with a clear role.

2. Example: A Simple Multi‑Agent Workflow

Imagine you want to publish a research‑backed article:

A Research Agent gathers sources and summaries
A Writing Agent drafts the article
An Editing Agent refines tone and clarity
A Publishing Agent formats and saves it into a capsule

Orchestration is the logic that connects these steps — passing outputs forward, managing state, and keeping everything in sync.

3. Defining a Flow

In Playnex, an orchestration flow is a defined sequence of agent steps.

From the CLI, you might create a new flow:

playnex flow create

You’ll define:

Name — what this flow is called
Steps — which agents run, and in what order
Inputs — what the flow needs to start
Outputs — where results are stored (often capsules)

4. Running a Flow

Once defined, you can run a flow from the CLI:

playnex flow run research-to-article

Playnex will:

Invoke each agent in sequence
Pass outputs from one step into the next
Store intermediate and final results into capsules

You can monitor progress from the CLI or the dashboard.

5. Shared Memory Between Agents

Orchestration is tightly connected to memory. Agents in a flow can read from and write to shared capsules.

For example:

The Research Agent writes a summary capsule
The Writing Agent reads that capsule to draft the article
The Editing Agent refines the draft and updates the capsule

This shared memory makes the flow feel coherent — like a team working in the same room.

6. When to Use Orchestration

Orchestration shines when:

Your task has clear stages
Different agents are better at different parts
You want repeatable, reliable workflows
You want to automate multi‑step processes