Integrate with Amazon S3

Set up a Pinecone storage integration with an Amazon S3 bucket using an IAM role to bulk import data into indexes and export audit logs.

This page shows you how to integrate Pinecone with an Amazon S3 bucket. Once your integration is set up, you can use it to import data from your Amazon S3 bucket into a Pinecone index hosted on AWS, or to export audit logs to your Amazon S3 bucket.

Before you begin

Ensure you have the following:

A Pinecone account.
An Amazon S3 bucket.

1. Create an IAM policy

In the AWS IAM console:

In the navigation pane, click Policies.
Click Create policy.
In Select a service section, select S3.
Select the following actions to allow:
- ListBucket: Permission to list some or all of the objects in an S3 bucket. Required for importing data and exporting audit logs.
- GetObject: Permission to retrieve objects from an S3 bucket. Required for importing data.
- PutObject: Permission to add an object to an S3 bucket. Required for exporting audit logs.
In the Resources section, select Specific.
For the bucket, specify the ARN of the bucket you created. For example: arn:aws:s3:::example-bucket-name
For the object, specify an object ARN as the target resource. For example: arn:aws:s3:::example-bucket-name/*
Click Next.
Specify the name of your policy. For example: "Pinecone-S3-Access".
Click Create policy.

Targeting a subdirectory (optional)

To write audit logs to a specific subdirectory within your S3 bucket (e.g., my-bucket/pinecone-logs/), you need to configure your IAM policy differently for ListBucket vs. object-level actions:

For ListBucket, use a Condition block with StringLike to specify the prefix. Include both the directory path with and without the trailing wildcard:

JSON

{
    "Sid": "ListBucketWithPrefix",
    "Effect": "Allow",
    "Action": "s3:ListBucket",
    "Resource": "arn:aws:s3:::example-bucket-name",
    "Condition": {
        "StringLike": {
            "s3:prefix": [
                "pinecone-logs/",
                "pinecone-logs/*"
            ]
        }
    }
}

For PutObject and GetObject, use the Resource specifier with the subdirectory path:

JSON

{
    "Sid": "ObjectActionsInSubdirectory",
    "Effect": "Allow",
    "Action": [
        "s3:PutObject",
        "s3:GetObject"
    ],
    "Resource": "arn:aws:s3:::example-bucket-name/pinecone-logs/*"
}

Complete example policy for subdirectory access:

JSON

{
    "Version": "2012-10-17",
    "Statement": [
        {
            "Sid": "ListBucketWithPrefix",
            "Effect": "Allow",
            "Action": "s3:ListBucket",
            "Resource": "arn:aws:s3:::example-bucket-name",
            "Condition": {
                "StringLike": {
                    "s3:prefix": [
                        "pinecone-logs/",
                        "pinecone-logs/*"
                    ]
                }
            }
        },
        {
            "Sid": "ObjectActionsInSubdirectory",
            "Effect": "Allow",
            "Action": [
                "s3:PutObject",
                "s3:GetObject"
            ],
            "Resource": "arn:aws:s3:::example-bucket-name/pinecone-logs/*"
        }
    ]
}

2. Set up access using an IAM role

In the AWS IAM console:

In the navigation pane, click Roles.
Click Create role.
In the Trusted entity type section, select AWS account.
Select Another AWS account.
Enter the Pinecone AWS VPC account ID: 713131977538
Click Next.
Select the policy you created.
Click Next.
Specify the role name. For example: "Pinecone".
Click Create role.
Click the role you created.
On the Summary page for the role, find the ARN.

For example: arn:aws:iam::123456789012:role/PineconeAccess
Copy the ARN.

You will need to enter the ARN into Pinecone later.

3. Add a storage integration

You can add a storage integration in the Pinecone console or with the API.

Use the console

In the Pinecone console, add an integration with Amazon S3.

Select your project.
Go to Manage > Storage integrations.
Click Add integration.
Enter a unique integration name.
Select Amazon S3.
Enter the ARN of the IAM role you created.
Click Add integration.

Use the API

Pass the ARN of your IAM role as the aws_iam_role.role_arn field:

curl

curl -sS -X POST "https://api.pinecone.io/storage-integrations" \
    -H "Api-Key: ${PINECONE_API_KEY}" \
    -H "X-Pinecone-Api-Version: unstable" \
    -H "Content-Type: application/json" \
    -d '{
      "name": "my-s3-integration",
      "provider": "s3",
      "aws_iam_role": {
        "role_arn": "arn:aws:iam::123456789012:role/pinecone-s3-access"
      }
    }'

Replace the role_arn value with the ARN of the IAM role you created, and my-s3-integration with a unique name for the integration.

The response includes the integration's id, which you need to import data, and a status of Validated or Invalid. If Pinecone can't assume the role, the request still succeeds and the integration is created with a status of Invalid, so check the status before you import.

Next steps

Import data from your Amazon S3 bucket into a Pinecone index.
Configure audit logs to export logs to your Amazon S3 bucket.