How BYOC works
Set up BYOC
Run workloads
Why use BYOC?
Bring-Your-Own-Cloud (BYOC) is ideal for Organizations with strict data residency, sovereignty, and compliance requirements.Data residency & sovereignty
Data residency & sovereignty
Data remains in your cloud
Data remains in your cloud
How BYOC works
When BYOC is enabled, Wherobots provisions a dedicated BYOC region for your Organization that maps to a region in your AWS account. Where a Hosted region runs your workloads on serverless compute in Wherobots’ account, a BYOC region runs them on infrastructure Wherobots manages inside your VPC. Workloads you target at this region run on compute inside your account.The BYOC region identifier
The BYOC region identifier
us-east-1 might appear as aws-us-east-1-abcd123ef.You’ll select this region when starting notebooks and pass it explicitly when submitting Job Runs or starting SQL Sessions.Storage in BYOC
Storage in BYOC
Data isolation
Data isolation
- Data processed in your BYOC region stays in your AWS account.
- A bucket whose trust policy allows only your BYOC account cannot be reached from Hosted regions.
Before you start
Prior to setting up and running workloads in BYOC, review the requirements for both Wherobots and AWS.Wherobots requirements
Wherobots requirements
- A Innovation or Enterprise Organization in Wherobots Cloud.
- BYOC enabled for your Organization. BYOC is provisioned by Wherobots — contact your account team or support@wherobots.com to enable it.
- An Admin account to set the Organization’s Default Region and create Storage Integrations.
- Members with the User role can run workloads in BYOC once BYOC has been configured. See Organization Roles.
AWS requirements
AWS requirements
- An AWS account in the region where your BYOC environment is provisioned.
- An S3 bucket for your S3 Storage Integration. BYOC has no access to Wherobots’ Managed Storage, so at least one Storage Integration is required to move data in and out.
- Sufficient Amazon EC2 (Elastic Compute Cloud) service quotas in the target region (Wherobots recommends a quota of at least
128 vCPUsforRunning On-Demand Standard A/C/D/H/I/M/R/T/Z instances).- New accounts often have low default quotas, which can block workload startup and leave sessions stuck in the
PENDINGstate.
- New accounts often have low default quotas, which can block workload startup and leave sessions stuck in the
- Sufficient Amazon EC2 (Elastic Compute Cloud) service quotas in the target region (Wherobots recommends a quota of at least
Set up BYOC
Setting up BYOC is a one-time configuration handled by an Organization Admin.What to provide, and when
To start BYOC provisioning, the customer only needs to provide three items:- Target AWS region
- AWS account ID (12 digits)
- Wherobots Organization ID
Process overview diagram
Process overview diagram
Request BYOC enablement
Provide details for provisioning
- Target AWS region: The region where you want your BYOC environment provisioned (for example,
us-east-1). - AWS account ID: The 12-digit account where the BYOC environment will be provisioned. Wherobots strongly recommends a new, dedicated AWS account; if that isn’t possible, tell the Wherobots contact handling your request.
- Wherobots Organization ID: Find it within Organization Settings.
Run CloudFormation
- Wherobots will provide you with the necessary CloudFormation templates and instructions to run in your AWS account. This sets up the infrastructure for your BYOC environment (including creating an IAM role and Role Actions), including the compute resources and permissions Wherobots needs to operate within your account. For more information about CloudFormation, see What is CloudFormation? in the AWS documentation.
-
Once you’ve been given the CloudFormation template, click Create stack in the AWS CloudFormation console.
This stack creates an IAM Role which will grant Wherobots permission to deploy infra into the BYOC account.
Once successfully created, the IAM role will appear as
CREATE_COMPLETEin the CloudFormation console.
Whitelist egress traffic
(Optional) Allow MCP access to your BYOC compute
us-west-2 region — see MCP & agentic tools.Wherobots needs inbound (ingress) access from its control plane to the compute it runs in your account. When users request the MCP server to execute SQL Queries, the MCP server opens a WebSocket connection to the SQL-session workload pod in the BYOC compute plane to relay queries and stream results between the customer’s agent and the SQL engine.To allow this, Wherobots will provide you with a list of IP address ranges to whitelist for inbound traffic. You will
add the Wherobots IP ranges to the relevant security groups so the control plane can reach the load balancer Wherobots provisions in front of those compute endpoints.Wherobots provisions your BYOC environment
- Amazon Virtual Private Cloud (VPC)
- Amazon Elastic Kubernetes Service (EKS)
- Amazon Elastic Compute Cloud (EC2) compute
- Amazon EC2 networking
- Elastic Load Balancing (ELB)
- AWS Identity and Access Management (IAM)
- AWS Key Management Service (KMS)
- Amazon CloudWatch Logs
- Amazon Route 53
- Amazon Simple Queue Service (SQS) and Amazon EventBridge
Confirm BYOC is enabled
- After Wherobots enables BYOC for your Organization, you can confirm your BYOC region is available.
- Log in to Wherobots Cloud.
-
Open Organization Settings → BYOC Regions
Each enabled region is listed with its Region ID and the AWS Account ID it runs in.

(Optional) Set BYOC as your Organization's Default Region (Admin)
- Go to Organization Settings → Preferences.
- Set Default Region to your BYOC region.
- Save your changes.
Create an S3 Storage Integration

Run workloads in BYOC
Once BYOC is set up, target the BYOC region from your Wherobots workloads. The core rule for BYOC compute is the same everywhere: point the workload at your BYOC region and read/write through an S3 Storage Integration.us-west-2 region, even when it executes your SQL against a BYOC runtime. See the MCP & agentic tools tab below.- Notebooks
- SQL Sessions
- Job Runs
- SDK & CLI
- MCP & agentic tools
- In Wherobots Cloud, click Start a Notebook.
- Set Region to your BYOC region. A BYOC-enabled region will be denoted with (BYOC) in the Start a Notebook dropdown. (If you set it as the Default Region, it’s already selected.)
- Launch the notebook and run your cells. Load and save data using your Storage Integration catalog.
Data isolation and access boundaries
BYOC is designed so that your data stays in your cloud. The following boundaries are enforced:| Data Relationship | Behavior |
|---|---|
| BYOC → Wherobots Managed Storage | BYOC regions cannot read from or write to Wherobots Managed Storage. |
| Hosted → restricted bucket | A bucket whose trust policy allows only your BYOC account cannot be accessed from Hosted regions. |
Next steps
After BYOC is set up, the next step is to create an S3 Storage Integration so you can read and write data in your BYOC workloads.Create a storage integration
Follow the instructions in the S3 Storage Integration guide to create an integration pointing at a bucket in your AWS account.Catalogs & Data Hub in BYOC
You can create and operate catalogs on your Storage Integration bucket from BYOC regions:- Create a catalog on your Storage Integration bucket in Data Hub, then run programmatic operations on that table as you would on any other table on the Wherobots Hosted platform. These operations work on BYOC the same way they do on the Wherobots Hosted platform. For more information, see Querying Datasets.
- Open Data catalogs are readable from BYOC workloads — your runtimes can read from Wherobots Open Data catalogs (such as Overture Maps) as usual.
Limitations
Current limitations
Current limitations
- The Wherobots MCP server is not yet deployed in BYOC regions; it runs in
us-west-2. Your SQL still executes against the runtime in your selected region (BYOC or Hosted), but the MCP server relays queries and routes responses back throughus-west-2. See Run workloads in BYOC. - The Default Region preference applies only to notebooks. Select the BYOC region explicitly for Job Runs and SQL Sessions.
- Standard S3 Storage Integration limitations apply (for example, bucket paths cannot contain periods, and a bucket can have a single storage integration).

