my-project-public

repository

loading code, commits, and activity

repositories

loading repo index

#1	---
#2	title: AWS Bedrock
#3	---
#4
#5	This integration demonstrates how to use Mem0 with AWS Bedrock and Amazon OpenSearch Service (AOSS) to enable persistent, semantic memory in intelligent agents.
#6
#7	## Overview
#8
#9	In this guide, you'll:
#10
#11	1. Configure AWS credentials to enable Bedrock and OpenSearch access
#12	2. Set up the Mem0 SDK to use Bedrock for embeddings and LLM
#13	3. Store and retrieve memories using OpenSearch as a vector store
#14	4. Build memory-aware applications with scalable cloud infrastructure
#15
#16	## Prerequisites
#17
#18	- AWS account with access to:
#19	- Bedrock foundation models (e.g., Titan, Claude)
#20	- OpenSearch Service with a configured domain
#21	- Python 3.8+
#22	- Valid AWS credentials (via environment or IAM role)
#23
#24	## Setup and Installation
#25
#26	Install required packages:
#27
#28	```bash
#29	pip install mem0ai boto3 opensearch-py
#30	```
#31
#32	Set environment variables.
#33
#34	Configure your AWS credentials using environment variables, IAM roles, or the AWS CLI.
#35
#36	```python
#37	import os
#38
#39	os.environ['AWS_REGION'] = 'us-west-2'
#40	os.environ['AWS_ACCESS_KEY_ID'] = 'AKIA...'
#41	os.environ['AWS_SECRET_ACCESS_KEY'] = 'AS...'
#42	```
#43
#44	## Initialize Mem0 Integration
#45
#46	Import necessary modules and configure Mem0:
#47
#48	```python
#49	import boto3
#50	from opensearchpy import OpenSearch, RequestsHttpConnection, AWSV4SignerAuth
#51	from mem0.memory.main import Memory
#52
#53	region = 'us-west-2'
#54	service = 'aoss'
#55	credentials = boto3.Session().get_credentials()
#56	auth = AWSV4SignerAuth(credentials, region, service)
#57
#58	config = {
#59	"embedder": {
#60	"provider": "aws_bedrock",
#61	"config": {
#62	"model": "amazon.titan-embed-text-v2:0"
#63	}
#64	},
#65	"llm": {
#66	"provider": "aws_bedrock",
#67	"config": {
#68	"model": "anthropic.claude-3-5-haiku-20241022-v1:0",
#69	"temperature": 0.1,
#70	"max_tokens": 2000
#71	}
#72	},
#73	"vector_store": {
#74	"provider": "opensearch",
#75	"config": {
#76	"collection_name": "mem0",
#77	"host": "your-opensearch-domain.us-west-2.es.amazonaws.com",
#78	"port": 443,
#79	"http_auth": auth,
#80	"embedding_model_dims": 1024,
#81	"connection_class": RequestsHttpConnection,
#82	"pool_maxsize": 20,
#83	"use_ssl": True,
#84	"verify_certs": True
#85	}
#86	}
#87	}
#88
#89	# Initialize memory system
#90	m = Memory.from_config(config)
#91	```
#92
#93	## Memory Operations
#94
#95	Use Mem0 with your Bedrock-powered LLM and OpenSearch storage backend:
#96
#97	```python
#98	# Store conversational context
#99	messages = [
#100	{"role": "user", "content": "I'm planning to watch a movie tonight. Any recommendations?"},
#101	{"role": "assistant", "content": "How about a thriller?"},
#102	{"role": "user", "content": "I prefer sci-fi."},
#103	{"role": "assistant", "content": "Noted! I'll suggest sci-fi movies next time."}
#104	]
#105
#106	m.add(messages, user_id="alice", metadata={"category": "movie_recommendations"})
#107
#108	# Search for memory
#109	relevant = m.search("What kind of movies does Alice like?", user_id="alice")
#110
#111	# Retrieve all user memories
#112	all_memories = m.get_all(user_id="alice")
#113	```
#114
#115	## Key Features
#116
#117	1. Serverless Memory Embeddings: Use Titan or other Bedrock models for fast, cloud-native embeddings
#118	2. Scalable Vector Search: Store and retrieve vectorized memories via OpenSearch
#119	3. Seamless AWS Auth: Uses AWS IAM or environment variables to securely authenticate
#120	4. User-specific Memory Spaces: Memories are isolated per user ID
#121	5. Persistent Memory Context: Maintain and recall history across sessions
#122
#123	<CardGroup cols={2}>
#124	<Card title="AWS Bedrock Cookbook" icon="aws" href="/cookbooks/integrations/aws-bedrock">
#125	Complete guide to using Bedrock with Mem0
#126	</Card>
#127	<Card title="Neptune Analytics Cookbook" icon="database" href="/cookbooks/integrations/neptune-analytics">
#128	Build graph memory with AWS Neptune
#129	</Card>
#130	</CardGroup>
#131
#132

z6Mkq5mY3JWtxoxUobWcfNHm7AkRubgSWEZTkBVqZXJviFZ5/my-project-public