Implementation:Microsoft Agent framework OpenAIResponsesClient Init

Knowledge Sources	Microsoft Agent Framework OpenAI API Reference
Domains	AI_Infrastructure, Agent_Architecture
Last Updated	2026-02-11 17:00 GMT

Overview

Concrete tool for configuring an OpenAI Responses API client provided by the Microsoft Agent Framework.

Description

The OpenAIResponsesClient class initializes an authenticated connection to the OpenAI Responses API. It extends multiple mixin layers providing middleware support, function invocation handling, and telemetry. The client implements the SupportsChatGetResponse protocol, making it compatible with the framework's Agent class and orchestration patterns.

Usage

Import this class when you need to connect to OpenAI's API directly (not through Azure). Configure with an API key (via parameter or environment variable OPENAI_API_KEY) and optionally specify a model ID. Use this client as the foundation for all agent operations.

Code Reference

Source Location

Repository: agent-framework
File: python/packages/core/agent_framework/openai/_responses_client.py
Lines: L1734-1846

Signature

class OpenAIResponsesClient(
    OpenAIConfigMixin,
    ChatMiddlewareLayer[OpenAIResponsesOptionsT],
    FunctionInvocationLayer[OpenAIResponsesOptionsT],
    ChatTelemetryLayer[OpenAIResponsesOptionsT],
    RawOpenAIResponsesClient[OpenAIResponsesOptionsT],
    Generic[OpenAIResponsesOptionsT],
):
    def __init__(
        self,
        *,
        model_id: str | None = None,
        api_key: str | Callable[[], str | Awaitable[str]] | None = None,
        org_id: str | None = None,
        base_url: str | None = None,
        default_headers: Mapping[str, str] | None = None,
        async_client: AsyncOpenAI | None = None,
        instruction_role: str | None = None,
        env_file_path: str | None = None,
        env_file_encoding: str | None = None,
        middleware: Sequence[ChatMiddleware | ChatMiddlewareCallable | FunctionMiddleware | FunctionMiddlewareCallable] | None = None,
        function_invocation_configuration: FunctionInvocationConfiguration | None = None,
        **kwargs: Any,
    ) -> None:

Import

from agent_framework.openai import OpenAIResponsesClient

I/O Contract

Inputs

Name	Type	Required	Description
model_id	str	No	OpenAI model name (e.g., "gpt-4o"). Also via env var OPENAI_RESPONSES_MODEL_ID.
api_key	str or Callable	No	API key or async provider. Also via env var OPENAI_API_KEY.
org_id	str	No	Organization ID. Also via env var OPENAI_ORG_ID.
base_url	str	No	Custom base URL. Also via env var OPENAI_BASE_URL.
default_headers	Mapping[str, str]	No	Default HTTP headers for requests.
async_client	AsyncOpenAI	No	Pre-configured OpenAI async client instance.
instruction_role	str	No	Role for instruction messages ("system" or "developer"). Default: "system".
middleware	Sequence	No	Chat and function middleware pipeline.
function_invocation_configuration	FunctionInvocationConfiguration	No	Override for function invocation behavior.

Outputs

Name	Type	Description
instance	OpenAIResponsesClient	Configured client implementing SupportsChatGetResponse protocol.

Usage Examples

Basic Initialization

from agent_framework.openai import OpenAIResponsesClient

# Uses OPENAI_API_KEY and OPENAI_RESPONSES_MODEL_ID env vars
client = OpenAIResponsesClient()

With Explicit Configuration

from agent_framework.openai import OpenAIResponsesClient

client = OpenAIResponsesClient(
    model_id="gpt-4o",
    api_key="sk-...",
    org_id="org-...",
)

With Middleware

from agent_framework.openai import OpenAIResponsesClient

# Custom middleware for security filtering
class SecurityMiddleware:
    async def __call__(self, context, next_handler):
        # Filter sensitive content
        return await next_handler(context)

client = OpenAIResponsesClient(
    middleware=[SecurityMiddleware()],
)

Related Pages

Implements Principle

Page Connections

Double-click a node to navigate. Hold to expand connections.

Principle

Implementation

Heuristic

Environment