Use Anthropic safely from the Assistant.
Claude models for long-context, instruction-following tasks. Route Anthropic model usage through governed Assistant workflows. Every read and write runs through a governed connection — scoped to what you allow, approved where it matters, and fully audited.
Capabilities
What the Assistant can do with Anthropic.
The top read and write actions, in plain language — each one a scoped action through a governed connection.
Example workflows
Put Anthropic to work.
Concrete ways teams use the Assistant with this tool — always inside the permission model.
Controls
Connected, but never unguarded.
Connecting Anthropic doesn't hand the Assistant the keys. Admins decide exactly what it can reach and do.
Connection detail preview
Anthropic governance controls
Enabled actions
Credential status
Stored in Vault, decrypted only at run time
Approval policy
Sensitive writes can require human sign-off
Audit trail
Every read, write, actor, and timestamp is logged
Govern Anthropic through Connections
Connect Anthropic once, set its scopes and approvals, and the Assistant can use it across the company — every action audit-logged.
See the Assistant use Anthropic safely.
Book a demo and we'll show the Assistant connected — and governed — on your own tools.