A decision is one payment request evaluated by the policy engine. Shadow mode is free and stays free. You pay when you turn enforcement on.
Every plan starts with shadow mode. Upgrade to stop bad refunds automatically and to put policy, approvals, and a sealed audit record on every other outflow.
See what a policy would have caught, before you pay anything
$0
or
Turn enforcement on
$299/mo
For the requirements that need a conversation
Custom
Break-even math.
Team costs $299 a month. It pays for itself the month it prevents that much in refunds you did not owe.
Monthly cost divided by your average prevented refund gives the number of refunds that has to catch.
Annual billing. Annual billing is available on request; contact us for terms.
Three inputs. One run-rate. No backend, no signup.
Three numbers in. One annual run-rate out. No backend, no signup — adjust live.
At 2,500 decisions/month, Axiru Team costs $299/mo and could prevent $5,313 in monthly leakage — a 17.8x return.
Estimate only. Actual prevented leakage depends on your policy and historical refund mix.
The full architecture is on /security; here’s the short version so you don’t have to leave pricing.
Your refund, credit, or adjustment code calls axiru.request_spend_approval first. Axiru evaluates policy and returns Allow, Deny, or Require Approval. The Stripe API call only fires on Allow. Wired in with ~10 lines via REST or MCP. This is the path that prevents bad outflows before money moves.
Read-only Stripe Connect feeds refund, dispute, payout, and transfer events into the decision ledger for shadow-mode replay and audit. Webhooks fire after Stripe has executed, so this path is observability: use it to evaluate Axiru on your own history before flipping enforcement on.
When we say decisions before money moves, we mean Path A. See the full architecture →
One definition. One price. Every rail you connect.
A live governed decision is any outbound value-transfer event evaluated by Axiru's policy engine with live enforcement active. The same definition applies across every rail you connect:
Auto-allowed, blocked, and approver-routed outcomes all count once, regardless of rail. Shadow-mode replays and historical imports are always free; they never count toward your decision volume.
Idempotent by design: agent retry loops will not bill you twice. Every POST /api/v1/decisions call accepts an idempotency_key in the body. The policy engine deduplicates by (workspace, idempotency_key) on a persistent store, so an agent that retries the same intent 50 times in a tight loop still results in one billed decision. Spend caps apply on top of dedupe: once you hit your cap, additional decisions either fail-closed (deny) or fail-open (allow + alert) based on your workspace policy.
Need policies that cite live external evidence (carrier tracking, Stripe disputes, merchant ToS, on-chain attestations) at decision time? Evidence Agent is available as a Control add-on and included at Enterprise.
Clear answers for self-serve onboarding.
Yes. Every new workspace begins in shadow mode. You only pay when you activate live enforcement.
Yes. Historical import is built into the free evaluation path. Shadow mode import does not count against decision volume.
No. You can start, evaluate, and run simulation entirely without contacting sales, Free Audit, Live Launch, and Control are all self-serve.
Connect Stripe at onboarding for live event ingestion, or upload historical data first. Read-only access only during shadow mode.
Yes. Simulation and replay happen before you turn on live enforcement, no production risk.
Read access to refund, dispute, payout, transfer, and customer balance events via Stripe webhook. We do not store payment method data. Connection is scoped to outflow governance events only.
Yes. Annual billing is available on request; contact us for terms. Switch at any time from your dashboard or contact us.
We replaced it with Live Launch ($199/mo), which is better for low-volume teams, and Control ($799/mo), which is the right fit once you're processing more than 2,500 decisions/month.
No. The same per-decision pricing applies to refunds on legacy Stripe, agent spend authorizations on Stripe DAA, and pre-authorization requests on x402. One definition of ‘governed decision’ spans every rail you connect. Cross-rail policies count exactly once even when one logical decision touches multiple rails.
Phase 1 today: Stripe (refund enforcement pre-execution; payouts, transfers, disputes, and platform fee refunds governed from webhooks), x402 (live, enforcing), Stripe DAA (shadow), and USDC on Solana (early access: pre-execution transfer decisions via the API). UCP, Tempo, Open USD (OUSD), USDC on Base, PYUSD, Plaid ACH, Modern Treasury, Dwolla, and Square are reserved in the type system; adapters ship as they’re available.
Yes. Live Launch and Control plans support a monthly spend cap. We notify you at 80% and again at 100%. You can adjust or remove the cap from your billing settings.
We discuss decision volume, multi-workspace setup, SOX controls, SSO requirements, and whether outcome-based pricing fits your model.
"Our team makes hundreds of refund, payout, and credit decisions a day. Some are human reps. Some may be automated. None of them are governed by a consistent policy with an audit trail. Axiru's shadow mode will show us, in dollars, what that costs across every outflow rail, before we spend anything on enforcement."
Start in shadow mode first. Move to live enforcement later.
We use cookies for analytics and marketing measurement. You can reject non-essential cookies at any time.