Item with id `rs_` not found

[shankar-dh]

Please read this first

• Have you read the docs? Agents SDK docs
• Have you searched for related issues? Others may have faced similar issues.

Describe the bug

In certain scenarios, the Agents SDK returns the following error:

Example:

openai.NotFoundError: Error code: 404 - {'error': {'message': "Item with id 'rs_3c4caff36e34c099006900f0e3c66481909c15ac262a15b060' not found.", 'type': 'invalid_request_error', 'param': 'input', 'code': None}}

Configuration:

• Store is not set to False
• Using default session management: SQLiteSession
• All items stored in the database are sent back to the agent for each turn
• Multi-agent system architecture with handoffs:
  • Agent 1: gpt-5 with minimal reasoning (acts as the triage agent)
  • Agent 2: gpt-5 with medium reasoning
  • Agent 3: gpt-5 with low reasoning

Issue:
Agent 1 generates an empty reasoning item as shown below:

8|f00b2f3b-770b-4f40-9ca3-ae7700940a19_6555bacd-f286-42b3-9f25-e7a9ae94e3de|{"id":"rs_3c4caff36e34c099006900f0e3c66481909c15ac262a15b060","summary":[],"type":"reasoning"}|2025-10-28 16:36:16

After several steps (including outputs produced by Agent 2 and Agent 3), when Agent 1 regains control and any new message is passed, the error above is triggered.

Debug information

• Agents SDK version: 0.2.8
• Python version: Python 3.11.6

Repro steps

While the code that triggers this error in our system involves multiple agents and handoffs, the following simplified code reproduces the same error:

import asyncio
import random

from agents import (
    Agent,
    ItemHelpers,
    Runner,
    enable_verbose_stdout_logging,
    function_tool,
)

enable_verbose_stdout_logging()


@function_tool
def how_many_jokes() -> int:
    """Return a random integer of jokes to tell between 1 and 10 (inclusive)."""
    return random.randint(1, 10)

ORIGINAL_RS_ID = "rs_07a11282cddc7074006905172505c88198bd41b4889a8d5649"

def mangle_rs_id(original_rs_id: str) -> str:
    return original_rs_id[:-2] + "ff"

async def main():
    agent = Agent(
        name="Joker",
        instructions="Call the `how_many_jokes` tool once, and tell that many jokes.",
        tools=[how_many_jokes],
        model="gpt-5-nano",
    )

    rs_id = mangle_rs_id(ORIGINAL_RS_ID)
    input_list = [
        {"content": "We want to hear jokes", "role": "user"},
        {
            "id": rs_id,
            "summary": [],
            "type": "reasoning",
        },
        {
            "arguments": "{}",
            "call_id": "call_xkkWToQpzfvn0asJfDPo06Wj",
            "name": "how_many_jokes",
            "type": "function_call",
            "id": "fc_07a11282cddc707400690517253a148198a2dad52c744c22ce",
            "status": "completed",
        },
        {
            "call_id": "call_xkkWToQpzfvn0asJfDPo06Wj",
            "output": "7",
            "type": "function_call_output",
        },
    ]

    result = Runner.run_streamed(
        agent,
        input=input_list,
    )
    async for event in result.stream_events():
        if (
            event.type == "run_item_stream_event"
            and event.item.type == "message_output_item"
        ):
            print(f"-- Message output:\n {ItemHelpers.text_message_output(event.item)}")


if __name__ == "__main__":
    asyncio.run(main())

Expected behavior

The SDK should not error out, as we are not modifying or altering any reasoning items.

[jcsalterego]

Important to note that when we were seeing this, we were providing a valid and recent reasoning item ID that was provided by a previous conversation, and the API was giving us a 404.

[ihower]

@jcsalterego I tried reproducing this issue using GPT-5 with three agents performing handoffs and streaming, as well as using SQLiteSession, just like you described, but I wasn’t able to trigger the same error.

Also, your repro code just intentionally mangles the reasoning item ID to force a 404, which produces the same error message but doesn’t actually help identify the real root cause.

[jcsalterego]

@ihower Thanks. Yes, the sample code reflects the scenario where we are using an ID provided by the responses API from an earlier call, but the API no longer recognizes the ID and returns a 404.

This behavior changed within the last week (Oct 24). This same code has been working for a while now, but it feels like the API has changed its retention behavior.

[jcsalterego]

@ihower Here is the exact error with the reasoning ID:

openai.NotFoundError: Error code: 404 - {'error': {'message': "Item with id 'rs_3c4caff36e34c099006900f0e3c66481909c15ac262a15b060' not found.", 'type': 'invalid_request_error', 'param': 'input', 'code': None}}

rs_3c4caff36e34c099006900f0e3c66481909c15ac262a15b060 is an item ID that was given to us from a previous Responses API call. This doesn't seem like a retention TTL issue because here are the timestamps for the conversation until it errored out:

2025-10-28 16:34:33
2025-10-28 16:34:33
2025-10-28 16:34:33
2025-10-28 16:35:40
2025-10-28 16:36:16
2025-10-28 16:36:16
2025-10-28 16:36:16
2025-10-28 16:36:16

Unfortunately, right now our only workaround is to monkey-patch OpenAIResponsesModel._fetch_response, look at the input when it's a list, and strip any id fields that are in the inner objects. It's fraught with peril, but there aren't many user-friendly options.

[ihower]

@jcsalterego I don’t see any possible cause other than something on the server side. But here’s a better workaround I can suggest: get the encrypted reasoning content (model_settings=ModelSettings(store=True, response_include=["reasoning.encrypted_content"] )) and include it in your next API call to OpenAI.
This way, the API should find the reasoning data, though it will cost you a bit more storage and bandwidth.