This script will download a file from a S3 bucket in one AWS account and stream it directly into another S3 bucket in another account. File is streamed as it arrives, so memory usage is low (typically <100MB)

s3-to-s3-cross-accounts.py

#! /bin/python

# This script will download a file from a S3 bucket in one AWS account
# and stream it directly into another S3 bucket in another account.
# File is streamed as it arrives, so memory usage is low (typically <100MB)

import boto3

# Create two S3 clients, one in each AWS account
# If wanting to use an instance role for one client, remove the access key parameters
s3_source = boto3.client(
    's3',
    aws_access_key_id='<Access Key Account A>',
    aws_secret_access_key='<Secret Access Key Account A>',
    aws_session_token='<Session Token Account A>'
)
s3_target = boto3.client(
    's3',
    aws_access_key_id='<Access Key Account B>',
    aws_secret_access_key='<Secret Access Key Account B>',
    aws_session_token='<Session Token Account B>'
)


# Change these to target the file you want
sourceBucket = '<Source S3 Bucket>'
sourceKey = '<Source Filename/Key>'
targetBucket = '<Target S3 Bucket>'
targetKey = '<Target Filename/Key>'


obj = s3_source.get_object( Bucket=sourceBucket, Key=sourceKey )
s3_target.upload_fileobj(obj['Body'], targetBucket, targetKey)