Agents
Every session, attributed
Every install and every session, with what it cost, by developer, model and agent. The line item has an owner.
- By developer, repository, model, agent
- Installs and sessions, not estimates
- A trend you can question
Use cases / AI Cost Governance
Use case
Every install and session, and what it cost, attributed to developers, repositories, models and agents, and Code Health for the same repositories. Bring your own provider key for the AI features that need one. It never leaves the platform.
Cost with an owner and a trend
Attributed to
What it cost, by developer, model and agent, next to code quality. The answer to the board question, without a data project.
Broken down, not one number
Cost and Code Health for the same repository
What Uzera shows finance and engineering
What each session cost, what it bought, and which key paid for it.
Agents
Every install and every session, with what it cost, by developer, model and agent. The line item has an owner.
Code Health
Cost by repository on one page, Code Health for the same repository on another. Put them side by side and you can see whether the spend bought better code.
AI provider
Bring your own Anthropic, OpenAI or Google key for the AI features that need one. It never leaves the platform, and the CLI on the developer's machine links no model SDK at all.
Agents
Re-runs are idempotent. The same commit is never scanned twice, so a retry never costs you twice.
How the cost gets an owner
Three steps, and the board question has an answer.
Install once. Every agent session on every developer machine is seen, with the model it used and the repository it touched.
Cost is attributed to the developer, the repository, the model and the agent. One number becomes a table with owners.
Cost sits next to Code Health, so the conversation is whether the spend bought better code, not where the number came from.
Where it fits
Before the budget conversation
Because spend arrives as one number with nothing behind it. Uzera attributes usage to developers, repositories and models, so the cost has an owner and a trend you can question.
Yours. Bring your own Anthropic, OpenAI or Google key for the AI features that need one, such as memory and convention discovery. Scanning and rule enforcement do not use it.
Cost is broken down by developer, repository and model, and Code Health scores the same repository scan after scan. Put the two side by side and the renewal conversation has both numbers.
Governance for your coding agents
AI writes your code. You stay in charge.