Mem0 Integration

Demonstrates using Mem0 as an external memory service for an Agno agent.

mem0_integration.py

"""
Mem0 Integration
================

Demonstrates using Mem0 as an external memory service for an Agno agent.
"""

from agno.agent import Agent, RunOutput
from agno.models.openai import OpenAIChat
from agno.utils.pprint import pprint_run_response

try:
    from mem0 import MemoryClient
except ImportError:
    raise ImportError(
        "mem0 is not installed. Please install it using `uv pip install mem0ai`."
    )


# ---------------------------------------------------------------------------
# Setup
# ---------------------------------------------------------------------------
client = MemoryClient()

user_id = "agno"
messages = [
    {"role": "user", "content": "My name is John Billings."},
    {"role": "user", "content": "I live in NYC."},
    {"role": "user", "content": "I'm going to a concert tomorrow."},
]

# Comment out the following line after running the script once
client.add(messages, user_id=user_id)


# ---------------------------------------------------------------------------
# Create Agent
# ---------------------------------------------------------------------------
agent = Agent(
    model=OpenAIChat(),
    dependencies={"memory": client.get_all(user_id=user_id)},
    add_dependencies_to_context=True,
)


# ---------------------------------------------------------------------------
# Run Example
# ---------------------------------------------------------------------------
if __name__ == "__main__":
    run: RunOutput = agent.run("What do you know about me?")
    pprint_run_response(run)

    input = [{"role": i.role, "content": str(i.content)} for i in (run.messages or [])]
    client.add(messages, user_id=user_id)

Current Mem0 API

Before running the copied recipe, replace its dependency lookup with the current filtered, paginated API:

agent = Agent(
    model=OpenAIChat(),
    dependencies={
        "memory": client.get_all(filters={"user_id": user_id}, page=1, page_size=50)["results"]
    },
    add_dependencies_to_context=True,
)

This loads one page of memories when the agent is constructed. Fetch subsequent pages if needed; refresh the dependency on later runs if the stored memories change. Mem0 processes writes asynchronously: wait for the seed write's event to finish before fetching its results. See the Mem0 migration guide for the event-status endpoint and pagination contract.

The final client.add(messages, ...) in the source repeats the seed data. To save this run's conversation instead, replace that final call with:

conversation = [
    {"role": message.role, "content": str(message.content)}
    for message in (run.messages or [])
    if message.role in {"user", "assistant"} and message.content
]
client.add(conversation, user_id=user_id)

Use a separate user ID for each user whose memories you store.

Run the Example

Set up your virtual environment

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

Install dependencies

uv pip install -U agno mem0ai openai

Export your API keys

export MEM0_API_KEY="your_mem0_api_key_here"
export OPENAI_API_KEY="your_openai_api_key_here"

Run the example

Save the code above as mem0_integration.py, then run:

python mem0_integration.py

Full source: cookbook/11_memory/integrations/mem0_integration.py