Implementation:Microsoft Agent framework BaseChatClient As Agent

Property	Value
Implementation Name	BaseChatClient As Agent
SDK	Microsoft Agent Framework
Repository	Microsoft Agent Framework
Source File	python/packages/core/agent_framework/_clients.py
Line Range	L438-517
Import	Access via any chat client instance (e.g., from agent_framework.openai import OpenAIChatClient)
Type	Factory method on BaseChatClient

Overview

The BaseChatClient.as_agent() method is a convenience factory that creates a fully configured Agent instance from an existing chat client. Rather than constructing an agent from scratch, callers invoke as_agent() on any chat client instance, passing specialization parameters such as name, instructions, and tools. The method wires the agent to the client's LLM connection, producing an agent ready for independent use or participation in multi-agent orchestration.

Code Reference

Source Location

Property	Value
File	python/packages/core/agent_framework/_clients.py
Method	as_agent()
Defined On	BaseChatClient
Lines	438-517

Signature

def as_agent(
    self,
    *,
    id: str | None = None,
    name: str | None = None,
    description: str | None = None,
    instructions: str = "",
    tools: FunctionTool | Callable | Sequence[FunctionTool | Callable] | None = None,
    default_options: OptionsCoT | None = None,
    chat_message_store_factory: ChatMessageStoreFactory | None = None,
    context_provider: ContextProvider | None = None,
    middleware: Sequence[...] | None = None,
    function_invocation_configuration: FunctionInvocationConfiguration | None = None,
    **kwargs: Any,
) -> Agent[OptionsCoT]:

Import Statement

from agent_framework.openai import OpenAIChatClient
# or
from agent_framework.azure import AzureOpenAIResponsesClient

I/O Contract

Inputs

Parameter	Type	Default	Description
id	None	None	Optional unique identifier for the agent.
name	None	None	The name of the agent, used for attribution in multi-agent conversations.
description	None	None	An optional human-readable description of the agent's purpose.
instructions	str	""	Role-specific system prompt that defines the agent's expertise domain and behavioral constraints.
tools	Callable | Sequence[FunctionTool | Callable] | None	None	Tools the agent can invoke. Accepts a single tool, a callable, or a sequence of tools/callables.
default_options	None	None	Default chat completion options (e.g., temperature, max tokens) for this agent.
chat_message_store_factory	None	None	Factory for creating the agent's conversation history store, enabling history isolation per agent.
context_provider	None	None	Provider for injecting additional context into the agent's requests.
middleware	None	None	Middleware pipeline applied to the agent's request/response cycle.
function_invocation_configuration	None	None	Configuration controlling how the agent invokes function tools.
**kwargs	Any		Additional keyword arguments forwarded to the Agent constructor.

Output

Type	Description
Agent[OptionsCoT]	A fully configured Agent instance that uses this chat client as its LLM backend. The agent is ready for independent invocation or participation in multi-agent orchestration patterns.

Usage Examples

Creating Specialized Agents

from agent_framework.azure import AzureOpenAIResponsesClient
from azure.identity import AzureCliCredential

client = AzureOpenAIResponsesClient(credential=AzureCliCredential())

# Create specialized agents
writer = client.as_agent(
    name="writer",
    instructions="You are a creative writer. Write engaging content.",
)
reviewer = client.as_agent(
    name="reviewer",
    instructions="You are a code reviewer. Provide constructive feedback.",
)

# Each agent can be used independently
response = await writer.run("Write a poem about AI")

Agent with Tools

from agent_framework.azure import AzureOpenAIResponsesClient
from azure.identity import AzureCliCredential

client = AzureOpenAIResponsesClient(credential=AzureCliCredential())

researcher = client.as_agent(
    name="researcher",
    instructions="You are a research assistant. Use the provided tools to find information.",
    tools=[web_search, document_lookup],
)

Related Pages

• Principle:Microsoft_Agent_framework_Agent_Specialization

Sources

• Repo | Microsoft Agent Framework