Dry run. Nothing is signed, nothing is spent.

Spend Policy Simulator

Before you hand an agent a budget, find out what that budget actually buys. This runs your proposed policy against the real, live prices of the endpoints your agent will call, using the same rules that enforce a payment session, and shows you exactly which calls settle, which get refused and why, and where the money runs out.

Start from

The policy you are considering

Budget (USD)

Max per call

Allowed hosts

Comma separated. Blank means any host is payable.

Expires in

Settles on

What the agent will call

Endpoints, one per line

https://api.example.com/data x12 repeats a call. @0.05 assumes a price rather than probing it.

No account needed. No wallet is touched, no session is created, and no credits move.

Nothing simulated yet

A preset is already loaded. Press the button and the platform will probe each endpoint for its live 402 price, then replay your policy over the result.

Create a session

How sessions work

Custody, authority, and execution: why the agent never holds a key.

Full walkthrough

From first session to a running agent, end to end.