Search User Memories

Search user memories using different retrieval methods such as last_n, first_n, and agentic retrieval.

memory_search.py
"""
Search User Memories
====================

This example shows how to search user memories using different retrieval
methods such as last_n, first_n, and agentic retrieval.
"""

from agno.db.postgres import PostgresDb
from agno.memory import MemoryManager, UserMemory
from agno.models.openai import OpenAIChat
from rich.pretty import pprint

# ---------------------------------------------------------------------------
# Setup
# ---------------------------------------------------------------------------
db_url = "postgresql+psycopg://ai:ai@localhost:5532/ai"
memory_db = PostgresDb(db_url=db_url)

# ---------------------------------------------------------------------------
# Create Memory Manager
# ---------------------------------------------------------------------------
memory = MemoryManager(model=OpenAIChat(id="gpt-5.6-luna"), db=memory_db)

# ---------------------------------------------------------------------------
# Run Memory Search
# ---------------------------------------------------------------------------
if __name__ == "__main__":
    john_doe_id = "john_doe@example.com"
    memory.add_user_memory(
        memory=UserMemory(memory="The user enjoys hiking in the mountains on weekends"),
        user_id=john_doe_id,
    )
    memory.add_user_memory(
        memory=UserMemory(
            memory="The user enjoys reading science fiction novels before bed"
        ),
        user_id=john_doe_id,
    )
    print("John Doe's memories:")
    pprint(memory.get_user_memories(user_id=john_doe_id))

    memories = memory.search_user_memories(
        user_id=john_doe_id, limit=1, retrieval_method="last_n"
    )
    print("\nJohn Doe's last_n memories:")
    pprint(memories)

    memories = memory.search_user_memories(
        user_id=john_doe_id, limit=1, retrieval_method="first_n"
    )
    print("\nJohn Doe's first_n memories:")
    pprint(memories)

    memories = memory.search_user_memories(
        user_id=john_doe_id,
        query="What does the user like to do on weekends?",
        retrieval_method="agentic",
    )
    print("\nJohn Doe's memories similar to the query (agentic):")
    pprint(memories)

last_n and first_n select stored memories by recency. agentic sends the user's stored memories and query to the configured model, which selects matching memory IDs. It does not use a vector index.

Run the Example

Set up your virtual environment

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

Install dependencies

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

Export your OpenAI API key

export OPENAI_API_KEY="your_openai_api_key_here"

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 the example

Save the code above as memory_search.py, then run:

python memory_search.py

Full source: cookbook/11_memory/memory_manager/04_memory_search.py