my-project-public

repository

loading code, commits, and activity

repositories

loading repo index

#1	---
#2	title: OpenAI Agents SDK
#3	---
#4
#5	Integrate [Mem0](https://github.com/mem0ai/mem0) with [OpenAI Agents SDK](https://github.com/openai/openai-agents-python), a lightweight framework for building multi-agent workflows. This integration enables agents to access persistent memory across conversations, enhancing context retention and personalization.
#6
#7	## Overview
#8
#9	1. Store and retrieve memories from Mem0 within OpenAI agents
#10	2. Multi-agent workflows with shared memory
#11	3. Retrieve relevant memories for past conversations
#12	4. Personalized responses based on user history
#13
#14	## Prerequisites
#15
#16	Before setting up Mem0 with OpenAI Agents SDK, ensure you have:
#17
#18	1. Installed the required packages:
#19	```bash
#20	pip install openai-agents mem0ai
#21	```
#22
#23	2. Valid API keys:
#24	- [Mem0 API Key](https://app.mem0.ai/dashboard/api-keys)
#25	- [OpenAI API Key](https://platform.openai.com/api-keys)
#26
#27	## Basic Integration Example
#28
#29	The following example demonstrates how to create an OpenAI agent with Mem0 memory integration:
#30
#31	```python
#32	import os
#33	from agents import Agent, Runner, function_tool
#34	from mem0 import MemoryClient
#35
#36	# Set up environment variables
#37	os.environ["OPENAI_API_KEY"] = "your-openai-api-key"
#38	os.environ["MEM0_API_KEY"] = "your-mem0-api-key"
#39
#40	# Initialize Mem0 client
#41	mem0 = MemoryClient()
#42
#43	# Define memory tools for the agent
#44	@function_tool
#45	def search_memory(query: str, user_id: str) -> str:
#46	"""Search through past conversations and memories"""
#47	memories = mem0.search(query, user_id=user_id, limit=3)
#48	if memories and memories.get('results'):
#49	return "\n".join([f"- {mem['memory']}" for mem in memories['results']])
#50	return "No relevant memories found."
#51
#52	@function_tool
#53	def save_memory(content: str, user_id: str) -> str:
#54	"""Save important information to memory"""
#55	mem0.add([{"role": "user", "content": content}], user_id=user_id)
#56	return "Information saved to memory."
#57
#58	# Create agent with memory capabilities
#59	agent = Agent(
#60	name="Personal Assistant",
#61	instructions="""You are a helpful personal assistant with memory capabilities.
#62	Use the search_memory tool to recall past conversations and user preferences.
#63	Use the save_memory tool to store important information about the user.
#64	Always personalize your responses based on available memory.""",
#65	tools=[search_memory, save_memory],
#66	model="gpt-4.1-nano-2025-04-14"
#67	)
#68
#69	def chat_with_agent(user_input: str, user_id: str) -> str:
#70	"""
#71	Handle user input with automatic memory integration.
#72
#73	Args:
#74	user_input: The user's message
#75	user_id: Unique identifier for the user
#76
#77	Returns:
#78	The agent's response
#79	"""
#80	# Run the agent (it will automatically use memory tools when needed)
#81	result = Runner.run_sync(agent, user_input)
#82
#83	return result.final_output
#84
#85	# Example usage
#86	if __name__ == "__main__":
#87
#88	# preferences will be saved in memory (using save_memory tool)
#89	response_1 = chat_with_agent(
#90	"I love Italian food and I'm planning a trip to Rome next month",
#91	user_id="alice"
#92	)
#93	print(response_1)
#94
#95	# memory will be retrieved using search_memory tool to answer the user query
#96	response_2 = chat_with_agent(
#97	"Give me some recommendations for food",
#98	user_id="alice"
#99	)
#100	print(response_2)
#101	```
#102
#103	## Multi-Agent Workflow with Handoffs
#104
#105	Create multiple specialized agents with proper handoffs and shared memory:
#106
#107	```python
#108	from agents import Agent, Runner, handoffs, function_tool
#109
#110	# Specialized agents
#111	travel_agent = Agent(
#112	name="Travel Planner",
#113	instructions="""You are a travel planning specialist. Use get_user_context to
#114	understand the user's travel preferences and history before making recommendations.
#115	After providing your response, use store_conversation to save important details.""",
#116	tools=[search_memory, save_memory],
#117	model="gpt-4.1-nano-2025-04-14"
#118	)
#119
#120	health_agent = Agent(
#121	name="Health Advisor",
#122	instructions="""You are a health and wellness advisor. Use get_user_context to
#123	understand the user's health goals and dietary preferences.
#124	After providing advice, use store_conversation to save relevant information.""",
#125	tools=[search_memory, save_memory],
#126	model="gpt-4.1-nano-2025-04-14"
#127	)
#128
#129	# Triage agent with handoffs
#130	triage_agent = Agent(
#131	name="Personal Assistant",
#132	instructions="""You are a helpful personal assistant that routes requests to specialists.
#133	For travel-related questions (trips, hotels, flights, destinations), hand off to the Travel Planner.
#134	For health-related questions (fitness, diet, wellness, exercise), hand off to the Health Advisor.
#135	For general questions, handle them directly using available tools.""",
#136	handoffs=[travel_agent, health_agent],
#137	model="gpt-4.1-nano-2025-04-14"
#138	)
#139
#140	def chat_with_handoffs(user_input: str, user_id: str) -> str:
#141	"""
#142	Handle user input with automatic agent handoffs and memory integration.
#143
#144	Args:
#145	user_input: The user's message
#146	user_id: Unique identifier for the user
#147
#148	Returns:
#149	The agent's response
#150	"""
#151	# Run the triage agent (it will automatically handoff when needed)
#152	result = Runner.run_sync(triage_agent, user_input)
#153
#154	# Store the original conversation in memory
#155	conversation = [
#156	{"role": "user", "content": user_input},
#157	{"role": "assistant", "content": result.final_output}
#158	]
#159	mem0.add(conversation, user_id=user_id)
#160
#161	return result.final_output
#162
#163	# Example usage
#164	response = chat_with_handoffs("Plan a healthy meal for my Italy trip", user_id="alex")
#165	print(response)
#166	```
#167
#168	## Quick Start Chat Interface
#169
#170	Simple interactive chat with memory:
#171
#172	```python
#173	def interactive_chat():
#174	"""Interactive chat interface with memory and handoffs"""
#175	user_id = input("Enter your user ID: ") or "demo_user"
#176	print(f"Chat started for user: {user_id}")
#177	print("Type 'quit' to exit\n")
#178
#179	while True:
#180	user_input = input("You: ")
#181	if user_input.lower() == 'quit':
#182	break
#183
#184	response = chat_with_handoffs(user_input, user_id)
#185	print(f"Assistant: {response}\n")
#186
#187	if __name__ == "__main__":
#188	interactive_chat()
#189	```
#190
#191	## Key Features
#192
#193	### 1. Automatic Memory Integration
#194	- Tool-Based Memory: Agents use function tools to search and save memories
#195	- Conversation Storage: All interactions are automatically stored
#196	- Context Retrieval: Agents can access relevant past conversations
#197
#198	### 2. Multi-Agent Memory Sharing
#199	- Shared Context: Multiple agents access the same memory store
#200	- Specialized Agents: Create domain-specific agents with shared memory
#201	- Seamless Handoffs: Agents maintain context across handoffs
#202
#203	### 3. Flexible Memory Operations
#204	- Retrieve Capabilities: Retrieve relevant memories from previous conversations
#205	- User Segmentation: Organize memories by user ID
#206	- Memory Management: Built-in tools for saving and retrieving information
#207
#208	## Configuration Options
#209
#210	Customize memory behavior:
#211
#212	```python
#213	# Configure memory search
#214	memories = mem0.search(
#215	query="travel preferences",
#216	user_id="alex",
#217	limit=5 # Number of memories to retrieve
#218	)
#219
#220	# Add metadata to memories
#221	mem0.add(
#222	messages=[{"role": "user", "content": "I prefer luxury hotels"}],
#223	user_id="alex",
#224	metadata={"category": "travel", "importance": "high"}
#225	)
#226	```
#227
#228	<CardGroup cols={2}>
#229	<Card title="OpenAI Tool Calls Cookbook" icon="wrench" href="/cookbooks/integrations/openai-tool-calls">
#230	Learn how to integrate Mem0 with OpenAI function calling
#231	</Card>
#232	<Card title="Agents SDK Tool Cookbook" icon="cube" href="/cookbooks/integrations/agents-sdk-tool">
#233	Build agents with OpenAI SDK tools
#234	</Card>
#235	</CardGroup>
#236
#237

z6Mkq5mY3JWtxoxUobWcfNHm7AkRubgSWEZTkBVqZXJviFZ5/my-project-public