> ## Documentation Index
> Fetch the complete documentation index at: https://agno-v2.mintlify.site/llms.txt
> Use this file to discover all available pages before exploring further.

# Memory

> Personalized memory and session summaries with vLLM.

```python memory.py theme={null}
"""
Personalized memory and session summaries with vLLM.
Prerequisites:
1. Start a Postgres + pgvector container (helper script is provided):
       ./cookbook/scripts/run_pgvector.sh
2. Install dependencies:
       uv pip install sqlalchemy 'psycopg[binary]' pgvector
3. Run a vLLM server (any open model).  Example with Phi-3:
       vllm serve microsoft/Phi-3-mini-128k-instruct \
         --dtype float32 \
         --enable-auto-tool-choice \
         --tool-call-parser pythonic
Then execute this script – it will remember facts you tell it and generate a
summary.
"""

from agno.agent import Agent
from agno.db.postgres import PostgresDb
from agno.models.vllm import VLLM
from rich.pretty import pprint

# ---------------------------------------------------------------------------
# Create Agent
# ---------------------------------------------------------------------------

# Change this if your Postgres container is running elsewhere
DB_URL = "postgresql+psycopg://ai:ai@localhost:5532/ai"

agent = Agent(
    model=VLLM(id="microsoft/Phi-3-mini-128k-instruct"),
    db=PostgresDb(db_url=DB_URL),
    update_memory_on_run=True,
    enable_session_summaries=True,
)

# -*- Share personal information
agent.print_response("My name is john billings?", stream=True)
# -*- Print memories and summary
if agent.db:
    pprint(agent.get_user_memories(user_id="test_user"))
    pprint(
        agent.get_session(session_id="test_session").summary  # type: ignore
    )

# -*- Share personal information
agent.print_response("I live in nyc?", stream=True)
# -*- Print memories and summary
if agent.db:
    pprint(agent.get_user_memories(user_id="test_user"))
    pprint(
        agent.get_session(session_id="test_session").summary  # type: ignore
    )

# -*- Share personal information
agent.print_response("I'm going to a concert tomorrow?", stream=True)
# -*- Print memories and summary
if agent.db:
    pprint(agent.get_user_memories(user_id="test_user"))
    pprint(
        agent.get_session(session_id="test_session").summary  # type: ignore
    )

# Ask about the conversation
agent.print_response(
    "What have we been talking about, do you know my name?", stream=True
)

# ---------------------------------------------------------------------------
# Run Agent
# ---------------------------------------------------------------------------

if __name__ == "__main__":
    pass
```

## Run the Example

<Snippet file="create-venv-step.mdx" />

**1. Install dependencies**

```bash theme={null}
uv pip install -U agno "psycopg[binary]" openai sqlalchemy
```

**2. Export your vLLM API key**

```bash Mac/Linux theme={null}
export VLLM_API_KEY="your_vllm_api_key_here"
```

```bash Windows theme={null}
$Env:VLLM_API_KEY="your_vllm_api_key_here"
```

<Snippet file="run-pgvector-step.mdx" />

**3. Install vLLM**

Install vLLM in the environment that will serve the model:

```bash theme={null}
uv pip install -U vllm
```

**4. Start vLLM**

Serve the model used by this example with tool calling enabled:

```bash theme={null}
vllm serve microsoft/Phi-3-mini-128k-instruct --dtype float32 --enable-auto-tool-choice --tool-call-parser pythonic
```

**5. Run the example**

Save the code above as `memory.py`, then run:

```bash theme={null}
python memory.py
```

Full source: [cookbook/90\_models/vllm/memory.py](https://github.com/agno-agi/agno/blob/v3.0.4/cookbook/90_models/vllm/memory.py)


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.