Multi-agent systems

Orchestrate multiple agents that collaborate on shared goals, hand off tasks to specialized peers, and coordinate through structured workflows. Multi-agent systems decompose complex problems into sub-tasks handled by purpose-built agents, using tool planning and inter-agent communication to produce coherent outcomes.

Overview

When to Use This Pattern

  • Your problem requires multiple specialized agents collaborating toward a shared goal

  • You need structured handoffs between agents with different capabilities or tool sets

  • Your system requires a supervisor or planner agent that delegates work to sub-agents

  • You want to build agentic pipelines where one agent’s output feeds into another

Akka Components Involved

  • Agents — individual specialized agents with distinct roles and tool sets

  • Workflows — orchestrate agent-to-agent handoffs and coordination logic

  • Session Memory — share context and state across collaborating agents

Sample Projects

The following sample projects demonstrate this pattern:

  • multi-agent — multiple agents coordinating on a shared task with structured handoffs

See Also