Add Members
Invite team members and configure their access to datasets and training projects in Datature Vi.
To invite a team member in Datature Vi, open Settings, go to Members, click + Add Member, fill in their email address and access settings, then click Invite Users. They receive an invitation email and gain access once they accept.
- You need Owner or Admin role to invite members
- Check your Collaborator balance before inviting: each invitation uses one slot whether or not it's been accepted (see Resource usage)
- Have the email addresses of the people you want to invite
Navigate to the Members tab
From your organization dashboard, click Settings at the bottom of the left sidebar and select the Members tab.
After sending invitations
Invited members appear in the Members table with a pending status until they accept. They count toward your Collaborator limit from the moment you invite them.
The new member appears in the Members table. Once they accept, their status updates and they can access the resources you assigned.
If a member hasn't accepted after some time, you can remove the pending invitation from the Members table to free the slot.
How access control works
Datature Vi uses two layers of access control. The organization role determines what a member can do across the workspace (manage settings, invite others, view billing). The resource-level role determines what they can do within specific datasets and training projects.
When you invite a member, you configure both layers: first the organization role, then which datasets and training projects they can access and at what level.
Organization roles
- Owner: Full control over the organization, including billing, plan changes, and member management. Each organization has one Owner.
- Admin: Can manage members, create resources, and configure settings. Cannot change billing or transfer ownership.
- Member: Can work within assigned datasets and training projects but cannot manage other members or organization settings.
Resource-level roles
For each dataset or training project you grant access to, you also set a role:
- Owner: Full control over that resource, including the ability to delete it.
- Editor: Can create, modify, and view content within the resource. Cannot delete the resource itself.
- Viewer: Read-only access. Can view assets, annotations, runs, and metrics but cannot make changes.
Common access patterns
Annotator (no training access): Set Organization Role to Member. Grant Editor access to the specific datasets they will annotate. Leave Training unselected. This gives them annotation access without visibility into training runs or models.
ML engineer (full project access): Set Organization Role to Admin or Member. Grant Owner or Editor access to All Datasets and All Trainings. This gives them full control over the data-to-training pipeline.
Read-only stakeholder: Set Organization Role to Member. Grant Viewer access to specific datasets and training projects. This lets them monitor progress and review metrics without risk of accidental changes.
Contractor (scoped, temporary): Set Organization Role to Member. Grant Editor or Viewer access to only the datasets and training projects for their engagement. Remove them when their contract ends to free the Collaborator slot.
For a written RACI-style map (who approves labels, exports, keys, and invoices) on top of these roles, see Roles and RACI checklist.
Frequently asked questions
Next steps
Updated 30 days ago