Files
agent-framework/python/packages/orchestrations
T
Ben Thomas c609b14f63 Python: Fix GroupChat orchestrator message cleanup issue (#3712)
* Fix GroupChat orchestrator message cleanup issue

Apply clean_conversation_for_handoff to GroupChatOrchestrator and
AgentBasedGroupChatOrchestrator _handle_response methods to remove
tool-related content that causes API errors from empty messages.

Fixes #3705

* Move orchestration related files to orchestrations package.

* Fix imports

---------

Co-authored-by: alliscode <bentho@microsoft.com>
Co-authored-by: Evan Mattson <evan.mattson@microsoft.com>
c609b14f63 ยท 2026-02-06 03:50:37 +00:00
History
..

Agent Framework Orchestrations

Orchestration patterns for Microsoft Agent Framework. This package provides high-level builders for common multi-agent workflow patterns.

Installation

pip install agent-framework-orchestrations

Orchestration Patterns

SequentialBuilder

Chain agents/executors in sequence, passing conversation context along:

from agent_framework_orchestrations import SequentialBuilder

workflow = SequentialBuilder().participants([agent1, agent2, agent3]).build()

ConcurrentBuilder

Fan-out to multiple agents in parallel, then aggregate results:

from agent_framework_orchestrations import ConcurrentBuilder

workflow = ConcurrentBuilder().participants([agent1, agent2, agent3]).build()

HandoffBuilder

Decentralized agent routing where agents decide handoff targets:

from agent_framework_orchestrations import HandoffBuilder

workflow = (
    HandoffBuilder()
    .participants([triage, billing, support])
    .with_start_agent(triage)
    .build()
)

GroupChatBuilder

Orchestrator-directed multi-agent conversations:

from agent_framework_orchestrations import GroupChatBuilder

workflow = (
    GroupChatBuilder()
    .with_orchestrator(selection_func=my_selector)
    .participants([agent1, agent2])
    .build()
)

MagenticBuilder

Sophisticated multi-agent orchestration using the Magentic One pattern:

from agent_framework_orchestrations import MagenticBuilder

workflow = (
    MagenticBuilder()
    .participants([researcher, writer, reviewer])
    .with_manager(agent=manager_agent)
    .build()
)

Usage with agent_framework

You can also import orchestrations through the main agent_framework package:

from agent_framework.orchestrations import SequentialBuilder, ConcurrentBuilder

Documentation

For more information, see the Agent Framework documentation.