GVNR answers one question before an AI agent acts: is this in budget, is it stuck in a loop, and is it allowed at all. Runs on your machine. No account, no hosted service, no payment rail.
A session fanning out into fifty subagents is paused inside the first minute, and you decide whether it carries on, not after four figures. On top of the rate sit a $20 cap per agent session, on by default, and day, week and month caps across the whole fleet, off until you set them.
A guard that only blocks gets switched off. The third answer is what makes it survivable: ordinary work passes silently, and the irreversible things stop and ask you.
In budget and on task. Nothing interrupts.
Over its spend cap, stuck in a loop, or reckless. Never executes.
Irreversible, or spending too fast. Pauses and asks you.
Deleting a tree costs nothing. Reading your credentials costs nothing. A budget check waves every one of them through on a full budget, so capability is judged separately.
Five projects for five clients, and nobody has to label a session. It reads the folder the work is happening in. Cap one client for the month and that one stops while the other four carry on.
Scoped how you need it: one agent, one tool, one team, or the whole fleet. It fires on spend velocity and repeat loops too, so the stop does not wait on a human noticing. The stop is itself receipted, naming who pulled it and when.
Each card is an agent running now: what it was asked to do, which model answered, and what it has cost. One of them just hit a rate limit.
normalise yesterday's events into the warehouse
compare our pricing against the top five competitors
pull every product listing from the supplier catalogue
work out why the payment webhook drops events
Hash-chained, naming the human the agent acted for, the tool it tried, the model that answered and the rule that decided. Edit one record and the chain breaks, and says which line.