Create a collaboration in Snowsight

You can create a new collaboration directly in Snowsight using a step-by-step wizard. The wizard guides you through configuring the collaboration name, adding collaborators, assigning roles, mapping resources, and reviewing the collaboration before creation.

Note

To create a collaboration, your role must have the CREATE COLLABORATION privilege. For more information, see Managing access to collaborations, resources, and data.

To create a collaboration:

  1. Sign in to Snowsight.

  2. In the navigation menu, select Data sharing » Data clean rooms.

  3. Select Create Collaboration.

The create collaboration wizard opens with four steps.

Step 1: Collaboration information

Enter the basic information for your collaboration:

  • Collaboration Name: A unique name for the collaboration.

  • Description: An optional description of the collaboration’s purpose.

To use Cortex Code to get suggestions for the collaboration name and description, select the Suggest button. Cortex Code analyzes the context and offers recommendations.

Select Next to proceed.

Step 2: Collaborator details

Configure the collaborators who will participate in the collaboration.

Note

Templates and data offerings must be registered in a registry before they can be added to a collaboration. You can register them using the Collaboration API or Cortex Code.

Owner collaborator

Your account is automatically added as the collaboration owner. You can configure the following for the owner:

  • Alias: A display name for your account in this collaboration.

  • Roles: Select the roles for the owner (Data Provider, Analysis Runner, or both).

  • Template IDs: Optionally specify template IDs to include from the owner’s account.

  • Data Offering IDs: Optionally specify data offering IDs to include from the owner’s account.

Additional collaborators

Select Add Collaborator to add other collaborators. For each additional collaborator, configure:

  • Alias: A display name for the collaborator.

  • Account Identifier: The Snowflake account identifier for the collaborator (for example, ORG_NAME.ACCOUNT_NAME).

  • Roles: Select the roles for the collaborator (Data Provider, Analysis Runner, or both).

  • Template IDs: Optionally specify template IDs expected from this collaborator.

  • Data Offering IDs: Optionally specify data offering IDs expected from this collaborator.

To use Cortex Code to get suggestions for collaborator roles and resource assignments, select the Suggest button.

Select Next to proceed.

Step 3: Resource mapping

Configure the access and execution rules for the collaboration. This step presents checkbox matrices that let you control:

  • Data offering access: Which analysis runners can access which data offerings.

  • Template access: Which analysis runners can use which templates.

  • Activation destinations: Which analysis runners can activate results to which destinations.

Select the appropriate checkboxes to grant access. To use Cortex Code to get suggestions for the run configuration, select the Suggest button.

Select Next to proceed.

Step 4: Final review

Review the complete collaboration configuration before submitting. The review step displays a read-only summary organized into tabs:

  • Basic Details: The collaboration name and description.

  • Collaborators: A table of all collaborators with their roles and status.

  • Data Offerings: The data offerings included in the collaboration.

  • Templates: The templates included in the collaboration.

  • Destinations: The activation destinations configured for analysis runners.

  • Raw Spec: The generated YAML specification for the collaboration.

To use Cortex Code to validate the collaboration specification before submitting, select the Validate button. Cortex Code reviews the spec and identifies potential issues.

When you’re satisfied with the configuration, select Submit. The collaboration is created and invitations are sent to the specified collaborators.

After creation

After you submit the collaboration:

  • A loading indicator shows the creation progress.

  • On success, you see a confirmation with a Next Steps card that provides guidance on what to do next, such as sharing templates or linking data offerings.

  • On failure, you see an error message with details. Select Debug to use Cortex Code to help diagnose the issue.

Auto-join

When creating a collaboration, you can optionally enable auto-join on the review step. When auto-join is enabled, you select a role and warehouse, and the owner’s account automatically joins the collaboration after creation without requiring a separate join step.

Note

Auto-join creates a Snowflake task that uses the selected warehouse to complete the join process. The task consumes compute credits while it runs. If your role isn’t SAMOOHA_APP_ROLE, it must have the EXECUTE TASK account-level privilege. For more information, see INITIALIZE.