Anthropic

Every Claude dollar, owned and measured.

The Claude Console reports usage by workspace, key and model. Valistry adds owners, outcomes and savings, across all your AI vendors.

Read-only · Usage, bills and settings · No prompts or replies by default

Overview · Anthropic

The Overview of a sample company cut to Anthropic: its AI cost, owners, opportunities and budgets.

Valistry × Anthropic

How much do we spend on AI?
The Claude Console gives youUsage and cost by workspace, API key and model.
Valistry addsOne total with every other AI vendor and seat.
Who owns each dollar?
The Claude Console gives youSpend by workspace and API key.
Valistry addsOwners by team, across workspaces and every vendor.
What did the spend produce?
The Claude Console gives youClaude Code usage per person.
Valistry addsCost per merge, and which Claude plan fits each developer.
Will we stay within budget?
The Claude Console gives youA spend limit on each workspace.
Valistry addsBudgets and alerts across all your AI, before spend passes them.
Where can we save?
The Claude Console gives youUsage by model.
Valistry addsWhere a cheaper model or caching saves, checked on your own runs.

Question	The Claude Console gives you	Valistry adds
How much do we spend on AI?	Usage and cost by workspace, API key and model.	One total with every other AI vendor and seat.
Who owns each dollar?	Spend by workspace and API key.	Owners by team, across workspaces and every vendor.
What did the spend produce?	Claude Code usage per person.	Cost per merge, and which Claude plan fits each developer.
Will we stay within budget?	A spend limit on each workspace.	Budgets and alerts across all your AI, before spend passes them.
Where can we save?	Usage by model.	Where a cheaper model or caching saves, checked on your own runs.

How we connect with Anthropic

Every endpoint, checked8endpoints readSee every endpoint

Claude API usage and cost

Usage and cost, by day, model and key. Claude Code use on the API, by person. Tokens, cache use, the bill, sessions, lines and commits. Never prompts, replies or names.

How to connect it

Claude Enterprise seats

Each member as a code, not a name. Each person's cost and tokens, by model. Claude Code commits, and the skills and connectors used. All but the members need read:analytics. Never conversations, prompts or project names.