Multi-Agent System Patterns

Explore runnable templates for orchestrating multiple agents with the GL AIP.

When to use this section: You need proven coordination patterns before customising them for production.

Audience: Engineers designing workflows and PMs creating workflows from requirements.

Use these examples to compare architectures (sequential, parallel, router, hierarchical, aggregator) and understand when to apply each one.

Prerequisites

• Python 3.10+

• Poetry installed

• This repository cloned locally:

  git clone [e-mail]:GDP-ADMIN/glaip-sdk.git

• Environment variables defined in .env (see .env.example):

  AIP_API_KEY=your-api-key-here
  AIP_API_URL=https://your-platform.com/api

Getting Started

From glaip-sdk/python/glaip-sdk:

poetry install
cp .env.example .env  # then edit with your credentials

Run any pattern example with Poetry, for example the sequential workflow:

poetry run python examples/multi-agent-system-patterns/sequential/sequential.py

Example Structure

1

Overview of when the pattern works best

A short description of the pattern and the scenarios it fits.

2

Demo scenario you can run immediately

A runnable script or example input that demonstrates the pattern.

3

Diagram showing agent relationships

A diagram (or link to a diagram) illustrating how agents interact.

4

Implementation steps with code snippets

Step-by-step instructions and example code for implementing the pattern.

5

Run commands and required environment variables

Commands to run the demo and the environment variables needed (e.g., API keys).

6

Sample output for validation

Expected output examples to verify the pattern is running correctly.

7

Notes and related documentation

Additional considerations, caveats, and links to deeper docs.

Pattern Library

Pattern	When to use	Example directory
Aggregator	Combine specialist insights into one briefing.	examples/multi-agent-system-patterns/aggregator/
Hierarchical	Delegate tasks through supervisors for complex workflows.	examples/multi-agent-system-patterns/hierarchical/
Parallel	Execute independent tasks simultaneously or compare variants.	examples/multi-agent-system-patterns/parallel/
Router	Direct each request to the right specialist.	examples/multi-agent-system-patterns/router/
Sequential	Refine answers step-by-step with predictable stages.	examples/multi-agent-system-patterns/sequential/

Looking for iterative loops? Explore examples/multi-agent-system-patterns/loop/ for an experimental variant.

Related Documentation

• Agents guide — Agent lifecycle, nesting, and runtime overrides.

• Tools guide — Upload Python tools and reuse catalog assets.

• Automation & scripting — Run pattern scripts inside CI pipelines or cron jobs.

• Security & privacy — Apply memory, PII, and artifact-sharing controls across agents.