Multi-agent patterns
Build multi-agent systems with message passing, shared memory, and coordination patterns.
Letta supports multi-agent coordination patterns that combine role-specific agents, tagging, and shared state via shared memory blocks.
Coordination styles
Section titled “Coordination styles”Most multi-agent systems in Letta use one or more of these coordination styles:
- Delegation: A coordinator routes work to specialized agents.
- Parallelization: Multiple agents work on parts of a task concurrently.
- Synthesis: A coordinator combines outputs into a final response.
These tutorials focus on architecture and coordination design, including role definitions, tag conventions, and shared memory strategy.
Coordination patterns
Section titled “Coordination patterns”Choose a pattern based on your use case:
Supervisor-worker Coordinate specialized agents under a supervisor with tag-based routing.
Parallel execution Multiple agents analyze the same task from different perspectives.
Round-robin Distribute tasks evenly across agents in a rotating sequence.
Producer-reviewer Iterative content refinement with feedback loops.
Hierarchical teams Multi-layer coordination with executives, managers, and workers.
Related resources
Section titled “Related resources”- Shared memory blocks - Share state between agents
- Memory blocks overview - Understanding core memory
- Shared memory tutorial - Step-by-step shared memory guide