Implementation:Openai Openai python Response MCP Call Failed

Knowledge Sources	Openai_Openai_python OpenAI API Reference
Domains	API_Types, Responses_API
Last Updated	2026-02-15 00:00 GMT

Overview

Concrete streaming event type for signaling that an MCP tool call has failed provided by the openai-python SDK.

Description

ResponseMcpCallFailedEvent is a Pydantic model emitted when an MCP tool call has failed. It carries the item_id identifying the MCP tool call item that failed, the output_index indicating which output item failed, and a sequence_number for event ordering. The type field is always "response.mcp_call.failed".

Usage

Import this type when building streaming event handlers that need to detect and handle MCP tool call failures, for example to log errors or retry the operation.

Code Reference

Source Location

• Repository: openai-python
• File: src/openai/types/responses/response_mcp_call_failed_event.py

Signature

class ResponseMcpCallFailedEvent(BaseModel):
    """Emitted when an MCP tool call has failed."""

    item_id: str
    output_index: int
    sequence_number: int
    type: Literal["response.mcp_call.failed"]

Import

from openai.types.responses import ResponseMcpCallFailedEvent

I/O Contract

Fields

Name	Type	Required	Description
item_id	str	Yes	The ID of the MCP tool call item that failed.
output_index	int	Yes	The index of the output item that failed.
sequence_number	int	Yes	The sequence number of this event.
type	Literal["response.mcp_call.failed"]	Yes	The type of the event. Always response.mcp_call.failed.

Usage Examples

import openai

client = openai.OpenAI()

stream = client.responses.create(
    model="gpt-4o",
    input="Use the MCP tool to fetch data.",
    stream=True,
)

for event in stream:
    if event.type == "response.mcp_call.failed":
        print(f"MCP call FAILED: item {event.item_id} "
              f"at output index {event.output_index}")

Related Pages

• Environment:Openai_Openai_python_Python_3_9_Plus