For every payment your AI agents make, ZenFix reads the public chain and confirms it really happened — the right amount, to the address you approved — before it’s marked done. A receipt anyone can re-check. It never holds your funds or keys.
Nothing leaves until it clears your rules. Each attempt runs through one deterministic pass and lands as an explainable record.
An agent submits who to pay, how much, and what for — a structured intent, not a raw transfer.
Budget, recipient, and your policies are evaluated in one deterministic pass — allow, hold, or block.
The outcome, its reason codes, and the full trail are written down. Nothing happens off the record.
Most tools take an agent’s word that a payment went through. ZenFix reads the chain and checks — before a Pay Run is marked done.
ZenFix never holds your funds or private keys. Your agent executes the payment itself and reports the transaction back.
On Base Sepolia, ZenFix reads the public chain and checks the transaction succeeded, moved USDC, and paid at least the amount you authorized.
Pin a payout address per merchant and verification requires the transfer to have gone there — a claim the chain doesn't back is rejected, not recorded.
The same decision language runs through the whole product — the list, the detail, and the audit trail.
Cleared every rule and executed inside budget.
Needs a human. Parked for review, nothing moved.
Violated a hard rule and was stopped cold.
ZenFix is the authorization and audit layer — deliberately never the custodian.
Every workspace is row-level isolated in the database. One tenant can never read another's Pay Runs, policy, or keys.
ZenFix decides and verifies — it never moves money. API keys are stored only as hashes; the full key is shown once and can only be revoked.
Each Pay Run keeps its intent, the rule-by-rule decision, the human review, and the execution proof — an append-only trail you can read end to end.
One click with Google. Free while in test mode on Base Sepolia — production access on request.