Mem0

Mem0 (pronounced “mem-zero”) enhances AI assistants and agents with an intelligent memory layer, enabling personalized AI interactions. It remembers user preferences and traits and continuously updates over time, making it ideal for applications like customer support chatbots and AI assistants.

Mem0 offers two powerful ways to leverage our technology: our managed platform and our open source solution.

If you're opening this Notebook on colab, you will probably need to install LlamaIndex 🦙.

python

%pip install llama-index llama-index-memory-mem0

Setup with Mem0 Platform

Set your Mem0 Platform API key as an environment variable. You can replace <your-mem0-api-key> with your actual API key:

Note: You can obtain your Mem0 Platform API key from the Mem0 Platform.

python

import os

os.environ["MEM0_API_KEY"] = "m0-..."

Using from_client (for Mem0 platform API):

python

from llama_index.memory.mem0 import Mem0Memory

context = {"user_id": "test_users_1"}
memory_from_client = Mem0Memory.from_client(
    context=context,
    api_key="m0-...",
    search_msg_limit=4,  # Default is 5
)

Mem0 Context is used to identify the user, agent or the conversation in the Mem0. It is required to be passed in the at least one of the fields in the Mem0Memory constructor.

search_msg_limit is optional, default is 5. It is the number of messages from the chat history to be used for memory retrieval from Mem0. More number of messages will result in more context being used for retrieval but will also increase the retrieval time and might result in some unwanted results.

Using from_config (for Mem0 OSS)

python

os.environ["OPENAI_API_KEY"] = "<your-api-key>"
config = {
    "vector_store": {
        "provider": "qdrant",
        "config": {
            "collection_name": "test_9",
            "host": "localhost",
            "port": 6333,
            "embedding_model_dims": 1536,  # Change this according to your local model's dimensions
        },
    },
    "llm": {
        "provider": "openai",
        "config": {
            "model": "gpt-4o",
            "temperature": 0.2,
            "max_tokens": 1500,
        },
    },
    "embedder": {
        "provider": "openai",
        "config": {"model": "text-embedding-3-small"},
    },
    "version": "v1.1",
}
memory_from_config = Mem0Memory.from_config(
    context=context,
    config=config,
    search_msg_limit=4,  # Default is 5
)

Initialize LLM

python

from llama_index.llms.openai import OpenAI

llm = OpenAI(model="gpt-4o", api_key="sk-...")

Mem0 for Function Calling Agents

Use Mem0 as memory for FunctionAgents.

Initialize Tools

python

def call_fn(name: str):
    """Call the provided name.
    Args:
        name: str (Name of the person)
    """
    print(f"Calling... {name}")


def email_fn(name: str):
    """Email the provided name.
    Args:
        name: str (Name of the person)
    """
    print(f"Emailing... {name}")

python

from llama_index.core.agent.workflow import FunctionAgent

agent = FunctionAgent(
    tools=[email_fn, call_fn],
    llm=llm,
)

python

response = await agent.run("Hi, My name is Mayank.", memory=memory_from_client)
print(str(response))

python

response = await agent.run(
    "My preferred way of communication would be Email.",
    memory=memory_from_client,
)
print(str(response))

python

response = await agent.run(
    "Send me an update of your product.", memory=memory_from_client
)
print(str(response))

Mem0 for ReAct Agents

Use Mem0 as memory for ReActAgent.

python

from llama_index.core.agent.workflow import ReActAgent

agent = ReActAgent(
    tools=[call_fn, email_fn],
    llm=llm,
)

python

response = await agent.run("Hi, My name is Mayank.", memory=memory_from_client)
print(str(response))

python

response = await agent.run(
    "My preferred way of communication would be Email.",
    memory=memory_from_client,
)
print(str(response))

python

response = await agent.run(
    "Send me an update of your product.", memory=memory_from_client
)
print(str(response))

python

response = await agent.run(
    "First call me and then communicate me requirements.",
    memory=memory_from_client,
)
print(str(response))