---
title: Persistent memory for OpenAI Agents SDK agents, on your own machine
slug: semurg-memory-for-openai-agents-sdk
date: 2026-10-25
author: Semurg
section: how-to
tags: [openai-agents, python, mcp, agents, memory, how-to]
summary: Give an OpenAI Agents SDK agent a memory that outlives the session, on your own disk, through the SDK's own streamable HTTP MCP server class, with the exact program our tests run.
status: published
covers: [route:POST /v1/mcp, route:POST /v1/memory/recall, mcp:memory_remember, mcp:memory_recall]
constants: [24.04, 3.12]
---
> Last checked on r11.132.20261027.a274a28 (27 Oct 2026): every command below ran on a fresh node of that release. The current release is r11.135.20261028.2ff2348 (28 Oct 2026), and the commands have not been re-checked on it yet.

> Tested end to end with `openai-agents 0.24.0` and `mcp 2.3.0`. The release it ran on is named in the checked line at the top of this page.

An agent built with the OpenAI Agents SDK forgets what it learned when its session ends. This guide gives it a memory that outlives the
session: a Semurg node on your own machine, reached through the SDK's own `MCPServerStreamableHttp`, where every fact stays on your disk
and comes back cited to where it was said. The memory is local: the program below calls no model and needs no OpenAI key.

> Diagram openai-agents-memory: Your Agents SDK program stores a fact through the SDK's streamable HTTP MCP server class; your node keeps it in the scope you named, on your own disk; a later run, another agent or a plain HTTP call on the same scope reads it back, cited.

## What you need

- An Intel or AMD x86_64 machine with Ubuntu 24.04, on Linux directly or on Windows through WSL2. One node is free, with no licence key.
- Python 3.12, which Ubuntu 24.04 ships. We test in the official `python:3.12-slim` container.

## Step one: install a node

On Ubuntu 24.04 (x86_64), with sudo:

```bash
curl -fsSL https://one.semurg.io/install | sudo bash
```

The [install page](/install) lists what the installer needs and what it changes. The installer ends by printing your API token, and the
node keeps a copy. Put it in a file only you can read, and have every new terminal load it:

```bash
mkdir -p ~/.config/semurg
(umask 077; sudo grep '^SEMURG_API_TOKEN=' /etc/semurg/semurg.env | cut -d= -f2- > ~/.config/semurg/token)
echo 'export SEMURG_API_TOKEN="$(cat ~/.config/semurg/token)"' >> ~/.bashrc
export SEMURG_API_TOKEN="$(cat ~/.config/semurg/token)"
```

## Step two: install the Agents SDK

Ubuntu keeps its own Python packages apart, so install into a virtual environment of your own:

```bash
sudo apt install -y python3-venv
python3 -m venv ~/semurg-openai-agents && . ~/semurg-openai-agents/bin/activate
pip install openai-agents
```

## Step three: connect

Save the program below as `snippet.py`. It is the program our tests run as written, shown here from the
[agents page](/agents#openai-agents) rather than copied, so what you read is always the file that ran.

```python
# Run: python snippet.py   (SEMURG_API_TOKEN must be set in the environment)
import asyncio, os
from agents.mcp import MCPServerStreamableHttp

URL = os.environ.get("SEMURG_MCP_URL", "http://127.0.0.1:4000/v1/mcp")
HEADERS = {"Authorization": "Bearer " + os.environ["SEMURG_API_TOKEN"],
           "X-Semurg-Scope": os.environ.get("SEMURG_MCP_SCOPE", "default")}
FACTS = {"entities": [{"name": "Ada", "type": "person"}, {"name": "Acme", "type": "organisation"}],
         "relations": [{"source": "Ada", "target": "Acme", "type": "WORKS_AT"}]}


async def main():
    async with MCPServerStreamableHttp(name="semurg", params={"url": URL, "headers": HEADERS, "timeout": 20},
                                       client_session_timeout_seconds=30) as server:
        await server.call_tool("memory_remember", {"facts": FACTS})
        print(await server.call_tool("memory_recall", {"subject": "Ada"}))


asyncio.run(main())
```

*Ran: this exact file, end to end against a fresh node on release r11.132.20261027.a274a28, with openai-agents 0.24.0, mcp 2.3.0. Install: `pip install openai-agents`. Its run record: https://one.semurg.io/agents#openai-agents*

It reads your token from `SEMURG_API_TOKEN`, the node's address from `SEMURG_MCP_URL` (without it, `http://127.0.0.1:4000/v1/mcp`) and
the scope from `SEMURG_MCP_SCOPE` (without it, the scope named `default`). It opens an `MCPServerStreamableHttp` named `semurg` with the
token and the `X-Semurg-Scope` header and a timeout on every call, stores two facts with `memory_remember` (Ada works at Acme) and prints
what `memory_recall` returns for Ada.

## Step four: run it

Give the program a scope of its own, so it never writes into a memory you use for real work:

```bash
export SEMURG_MCP_SCOPE=try-openai-agents
python snippet.py
```

It prints the tool result the SDK hands back, which carries the node's answer for Ada: `state` set to `answered` and the `WORKS_AT` fact
that links Ada to Acme, each item cited. Storing the same facts again changes nothing, so it is safe to run twice.

Now read the same fact from outside the SDK, over plain HTTP. What one client writes, any other client on the same node and scope reads:

```bash
curl -sS http://127.0.0.1:4000/v1/memory/recall \
  -H "Authorization: Bearer $SEMURG_API_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{"scope":"try-openai-agents","subject":"Ada"}'
```

## The memory tools in the Agents SDK

The program calls each tool by name with `server.call_tool`. To let an agent choose, pass the same server to it as
`mcp_servers=[server]`; the agent then sees every memory tool under its own name (our test calls the tools directly, without a model).

- `memory_status`: reports whether shared memory is on for this node and which of your scopes hold facts.
- `memory_remember`: stores entities, the relationships between them and the passages they came from, or plain text, with an optional source label and date.
- `memory_recall`: recalls what is stored about one named entity, its own facts first, each item cited.
- `memory_relate`: adds one typed relationship between two entities in a scope, creating an end that is not stored yet.
- `memory_context`: answers a question in plain words with cited items and a verdict on whether the store covers it.
- `memory_document`: takes a whole document (PDF with a text layer, Word, HTML or plain text) and stores it as cited passages in a scope.
- `memory_forget`: deletes one thing per call: an entity and everything that names it, one fact, one document, every document under a source label, or the whole scope. It is unreachable at once and not yet wiped from the disk.
- `memory_export`: reads everything a scope holds, one page per call: send each reply's `next` back as `cursor` until it is empty. The records are the entities, the facts with their dated sources and whether each is current or ended, and each stored document once; a forgotten passage is never exported.
- `memory_import`: stores the records `memory_export` produced into any scope, through the same writes `memory_remember` makes. A page sent twice stores nothing twice, and the reply names every record it skipped.

Every answer says its `state`: `answered`, `nothing_stored` (nothing stored matched, so the agent should say it does not know) or `could_not_judge`.

## Troubleshooting

- `KeyError: 'SEMURG_API_TOKEN'`: the token is not in this shell. Run the last line of step one, or open a new terminal.
- A refused token can come back as a general error such as `MCPError: Server returned an error response`, without the reason. Check the token first: read it again from the node as in step one.
- After the node restarts (an update restarts it), an open `MCPServerStreamableHttp` did not recover by itself in our October 2026 restart test, run on an earlier release. Leave the `async with` block and open a new one; a new one answers at once.
- Keep the two timeouts the program sets (`timeout` in the parameters and `client_session_timeout_seconds`): they bound how long a call waits, so a stalled node shows up as a timeout error instead of a hang.
- `error: externally-managed-environment` from pip: you are outside the virtual environment. Activate it with `. ~/semurg-openai-agents/bin/activate`.

Use `127.0.0.1` rather than `localhost`: on Ubuntu `localhost` can resolve to the IPv6 address first, and the node listens on IPv4 loopback
only. To reach a node on another machine, set `SEMURG_MCP_URL`, or forward its port with `ssh -N -L 4000:127.0.0.1:4000 you@your-node`.

## Next

- The [agents page](/agents): every framework program, with its versions and the release it last ran on in our tests.
- [Give your agent persistent memory with MCP in five minutes](/blog/agent-memory-mcp-five-minutes): the memory calls over plain HTTP too.
- [Connecting to Semurg over MCP](/blog/connect-over-mcp): the endpoint, the protocol versions it speaks and how a refusal comes back.
- If a step does not work, tell us on the [contact page](/contact), with the reply you got.

*This post was written with AI assistance. Every result in it is checked against a published release or record before the post goes out.*