Custom Memory Manager

The local database setup below requires Docker.

Pass additional_instructions to append rules to the default memory-manager prompt. Pass system_message to replace the prompt.

Code

custom_memory_manager.py

from agno.agent import Agent
from agno.db.postgres import PostgresDb
from agno.memory import MemoryManager
from agno.models.openai import OpenAIResponses
from rich.pretty import pprint

db_url = "postgresql+psycopg://ai:ai@localhost:5532/ai"

db = PostgresDb(db_url=db_url)

memory_manager = MemoryManager(
    model=OpenAIResponses(id="gpt-5.2"),
    additional_instructions="""
    IMPORTANT: Don't store any memories about the user's name. Just say "The User" instead of referencing the user's name.
    """,
    db=db,
)

john_doe_id = "john_doe@example.com"

agent = Agent(
    model=OpenAIResponses(id="gpt-5.2"),
    db=db,
    memory_manager=memory_manager,
    update_memory_on_run=True,
    user_id=john_doe_id,
)

agent.print_response(
    "My name is John Doe and I like to swim and play soccer.", stream=True
)

agent.print_response("I don't like to swim", stream=True)


memories = agent.get_user_memories(user_id=john_doe_id)

print("John Doe's memories:")
pprint(memories)

Usage

Set up your virtual environment

uv venv --python 3.12
source .venv/bin/activate

Set your API key

export OPENAI_API_KEY=xxx

Install dependencies

uv pip install -U agno openai sqlalchemy 'psycopg[binary]'

Run PgVector

docker run -d \
  -e POSTGRES_DB=ai \
  -e POSTGRES_USER=ai \
  -e POSTGRES_PASSWORD=ai \
  -e PGDATA=/var/lib/postgresql \
  -v pgvolume:/var/lib/postgresql \
  -p 5532:5432 \
  --name pgvector \
  agnohq/pgvector:18

Run example

python custom_memory_manager.py