Skip to main content
This feature is available on Standard and Enterprise plans.
Exporting namespaces from a backup is the recommended way to rearchitect your use case and perform large-scale data revisions. It is faster and cheaper than using other API calls.

​
What you can do

  • Export a full backup or selected namespaces to your Amazon S3 or Google Cloud Storage bucket.
    Azure Blob Storage support is coming soon.
  • Handle the exported data accordingly (e.g., add metadata fields, split into more namespaces).
  • The exported data is already formatted for import. For a seamless import, maintain the existing Parquet schema when handling your data.

​
Before you export

Ensure you have a storage integration and a backup of the index you want to export.

​
Add a storage integration

Create an integration that allows Pinecone to write exported data to your object storage.
Storage integrations are scoped to the Project level. Be sure your storage integration is created within the Project in which your target index resides.
You’ll need a storage integration with read permissions to import your data back into Pinecone. We recommend also creating it at this time.
Follow Integrate with Amazon S3 to create an IAM policy, IAM role, and storage integration.For export, the IAM policy must allow s3:ListBucket and s3:PutObject on your destination bucket or subdirectory. For example:
{
  "Version": "2012-10-17",
  "Statement": [
    {
      "Effect": "Allow",
      "Action": [
        "s3:ListBucket",
        "s3:PutObject"
      ],
      "Resource": [
        "arn:aws:s3:::BUCKET_NAME",
        "arn:aws:s3:::BUCKET_NAME/*"
      ]
    }
  ]
}
When creating the IAM role, use a trust policy that allows Pinecone’s AWS account to assume the role. For example:
{
  "Version": "2012-10-17",
  "Statement": [
    {
      "Effect": "Allow",
      "Principal": {
        "AWS": [
          "arn:aws:iam::713131977538:root"
        ]
      },
      "Action": "sts:AssumeRole",
      "Condition": {}
    }
  ]
}

​
Create a backup

Create a backup of the serverless index you want to export. Then list the backups for the index and copy the backup_id of the backup to export.

​
Request an export

To start an export, contact Support and share the following:
  • Destination URI — The full URI of the location in your bucket where the exported data should be written. For example:
    • Amazon S3: s3://BUCKET_NAME/EXPORT_DIR
    • Google Cloud Storage: gs://BUCKET_NAME/EXPORT_DIR
  • Storage integration ID — The ID of the storage integration you created, as found in the Pinecone console.
  • Project ID — The ID of the project containing the backup to export, as found in the Pinecone console.
  • Backup ID — The ID of the backup to export.
  • Namespaces (optional) — The namespace name(s) to export, as found in the Pinecone console or via list_namespaces. Only needed if exporting specific namespaces.
Once Pinecone Support receives this information, the export operation will begin and you’ll be periodically updated on its status.