Govern AI Spend

See AI Spend, Attribute It, Then Cap It

Netra meters tokens from Claude Code, Claude API, OpenAI API, Codex, Gemini CLI, and more, ties each one to a user, department, model, and use case, prices it at your rates, and enforces quotas before the invoice arrives.

Get a Demo

The Challenge

One AI Invoice, No Owner

The invoice says how much. It doesn’t say who, or for what.

01

One Invoice, Many Owners

Vendor bills arrive as totals. They don’t show which department, user, or use case created the spend, so nobody owns it.

02

Agents Outspend People

One autonomous agent run can outspend a week of chat. Seat counts no longer predict the bill.

03

Overruns Show Up Late

Without a cap, a budget overrun appears when the invoice does. By then, the tokens are spent.

See and Attribute

From One Bill to Every User, Model, and Use Case

Measured at the endpoint and Gateway. No SDK and no application code changes.

01Overview

Tokens, Cost, and Broken Channels at a Glance

Overview shows total tokens and cost in USD, input and output tokens, active users, top users by token usage, top models by cost, and quota status. Broken Channels lists any metered channel that stopped reporting, so a data gap is never mistaken for a drop in spend.

02Token Insights

Every Token Has a Name and a Purpose

Token Usage breaks tokens down by user, department, model, and channel, splits prompt from completion tokens, and shows the topic they were spent on. Users and departments sync from Okta, Microsoft Entra ID, Google Workspace, or LDAP/AD, so there is no manual tagging.

03AI Usage Insights

Chat, Assisted, or Autonomous: Where Spend Goes

Conversations are grouped into use-case scenarios automatically. Agentic Activity splits human-led chat from tool-assisted work and autonomous agent runs, by tool and department. Account Activity shows where AI runs on corporate accounts and where on personal ones, and policies can treat each account type differently.

04Cost Table

Your Rates, Not List Price

Enter the input and output price per 1M tokens you actually pay for each model. Netra uses your Cost Table to turn tokens into USD across Token Insights and for quotas set in dollars. Finance sees figures it recognizes, and it is a quick setup step during a proof of value.

Cap

Set the Ceiling Before the Invoice Arrives

Token Quota turns a budget into a limit Netra enforces per channel and model.

AI Chargeback by Department

Start With Showback, Then Charge Back

Budget owners rarely accept a charge they can’t check. Netra gives finance the detail behind each figure: who used which tool, on which model, for what, at your rates. Share it as showback first, and move to chargeback once the numbers hold up.

  • Users and departments synced from Okta, Microsoft Entra ID, Google Workspace, or LDAP/AD
  • Cost in USD from your own Cost Table, not list price
  • AI Adoption by Department puts sessions per user next to the spend
  • Copilot Credits brings in Copilot Studio agent consumption via the Microsoft Power Platform API, attributed to departments
  • Export Report from Token Insights to share the numbers

Next Steps

Spend Is One Part of AI Governance

The same endpoint agent and Gateway cover AI risk and data protection.

Netra Token Governance

The product behind this page: Token Insights, AI Usage Insights, Token Quota, and Quota Alerts.

Learn more

Secure AI Coding

Control coding agents like Claude Code, Cursor, and Codex on developer machines.

Learn more

Discover Shadow AI

Find the AI tools, agents, and personal accounts in use across your endpoints.

Learn more

Stop AI Data Leakage

Detect and block sensitive data in prompts, uploads, and responses.

Learn more

FAQ

Common Questions

How does Netra support AI chargeback by department?

Netra attributes each metered token to a user and a department synced from your identity provider, such as Okta, Microsoft Entra ID, Google Workspace, or LDAP/AD. Your Cost Table converts tokens to USD at your own rates. You can share department totals as showback first, then move to chargeback once budget owners trust the numbers.

Which AI tools does Netra meter tokens for?

Netra meters tokens for Claude Code, Claude API, Claude Cowork, OpenAI API, Codex Desktop, Codex CLI, Kiro CLI, Kiro IDE, Antigravity IDE, Gemini CLI, and more. Copilot Studio agent consumption is tracked separately as Copilot Credits.

Can Netra stop AI spending when a quota runs out?

Yes. Each Token Quota sets a limit in tokens or USD for a channel and model, tenant-wide or for one department, over a calendar month, calendar week, or custom date range. It warns at 80% and alerts at 100% by default. With On Exceeded set to Circuit Break, Netra blocks the channel once the quota is exceeded, either for that department only or tenant-wide.

Does Netra need an SDK or code changes?

No. A standard SaaS deployment only requires the endpoint agent, storage setup, and users and departments synced from your IdP.

Does Netra track Microsoft Copilot Studio spend?

Yes. Token Insights includes a Copilot Credits view that pulls Copilot Studio agent consumption through the Microsoft Power Platform API and attributes it to departments, alongside your API and CLI token usage.

Put a Name on Every AI Dollar

Run a proof of value on your own endpoints and see tokens and cost by user, department, model, and use case.

Get a Demo