Authorization — before the action
Liability paperwork for the agentic economy.
Before your AI agent touches money or consequential actions, record the authorization: who approved what, for which agent, at which time, with what declared scope — sealed into a SHA-256 hash-chained ledger you can export if there's ever a dispute.
Issue your first receipt — free See pricing
Banks are asking who is liable when an agent acts. Regulators admit new legislation is needed. Your defense starts with a paper trail.
How it works
- Issue a receipt before the transaction. Approver, agent name + version + operator, action description, amount + currency, declared scope (the actions this authorization covers), expiry.
- It seals into a hash-chained ledger. Each receipt hashes the previous receipt's hash. Tamper with any entry and chain verification fails.
- Export a dispute pack if needed. JSON or printable HTML evidence exports — Pro and Team plans.
What receipts are (and aren't)
- Sensitive identifiers (account numbers) are stored as one-way hashes only — never plaintext.
- Every receipt carries a UTC timestamp and the full agent identity.
- Scope is declared explicitly: only the listed actions are covered by the authorization.
Honest scope: Evidence supporting disputes — not legal protection or legal advice.
Pricing
Free for up to 10 receipts/month. Pro $29/mo adds unlimited receipts, ledger export and dispute packs. Team $79/mo adds 5 seats and a shared ledger.