Home / Glossary / Multi-agent
What is Multi-Agent System?
A multi-agent system uses multiple specialised agents that delegate sub-tasks to each other. osFoundry’s agents call other agents as typed tools, enabling planner / executor / reviewer pipelines.
Detail
Instead of one agent trying to do everything, multi-agent systems decompose: a planner decides steps, executors specialised by domain do them, a reviewer verifies. Each agent has its own scope, prompt, and tools.
Multi-agent shines when sub-tasks need different expertise (e.g. one agent for web research, one for code, one for writing). It also enables better cost control — use cheap models for simple delegation, expensive models only where needed.
How osFoundry approaches Multi-Agent System
osFoundry exposes sub-agent calls as typed tools. Each agent has its own profile (system prompt, tools, memory, model). Multi-agent pipelines are built the same way as any tool call.
Related terms
Related features