What goes wrong
Your AI editor shows a usage number. That's a rough check, not an answer. After a real week of work you still don't know which model took the money, whether autocomplete ran all day in the background, or whether agents cost more than chat.
So “we use AI a lot” becomes the only explanation for a rising bill — and the only fix is guesswork.
Why it matters
AI spend is not one number. Autocomplete can drip all day. Agents can spike in an afternoon. Switching models can change the price without the editor feeling any busier. If you only watch the monthly total, you keep cutting the wrong thing.
What good looks like
In a few minutes you can say which models ran, what mix of chat, autocomplete, and agents you actually used, and when usage jumped — without sending the contents of those chats to another company's logger.
How Aura Wire helps
Aura Wire sits between your editor and the AI services it calls. It records what happened — which service, which model, how long, roughly how much — never the prompts or answers.
- Live charts for chat vs autocomplete vs agents
- Cost views based on those estimates, not a single vendor screen
- A timeline so quiet hours and busy agent sessions are obvious
Try Aura Wire free
AI chat for VS Code with a local pipeline behind it — live spend, secrets stopped, nothing stored.