Pricing
Scoring against your own history is free and always will be. What you pay for is the part your own ledger structurally cannot produce: what a merchant looks like across every account, and the evidence trail for when a payment goes wrong.
Billing is not configured on this deployment, so paid plans cannot be purchased here. The free tier, the API and the MCP server all work.
Free
One agent's worth of scoring, against your own history. No card.
- 50 risk checks per rolling day
- Spend caps: per agent, per merchant, per category, per rolling window
- Merchant and category allow / block lists
- Amount-anomaly, velocity, retry-loop and first-seen-merchant scoring against your own ledger
- x402 and AP2 protocol checks: quote overrun, asset mismatch, nonce replay, mandate scope and expiry
- Read-only transaction log of every decision
- Stateless mode: score against a ledger you send, no account needed
- Public OpenAPI 3.1 spec, llms.txt and MCP server
- Cross-agent and cross-merchant graph scoring
- Chargeback-assist claims and evidence packets
Pro
then $0.01 per risk check, metered
Scoring against the whole graph, not just your own history, plus the evidence trail for when a payment goes wrong.
- 50,000 risk checks a month included, then $0.01 each, metered
- Cross-agent and cross-merchant graph scoring: merchant fan-in, platform-wide denial rates, agent fan-out and coordinated bursts across unrelated accounts
- Confirmed-bad-counterparty signal from claims raised across the platform
- Chargeback-assist: raise a claim against a scored payment and get an evidence packet, PDF or JSON, with the decision, its signals and the surrounding activity
- API keys with per-key usage attribution
- Everything in Free
Scale
Above 2 million checks a month, private signal overlays, or a deployment inside your own perimeter. Priced per engagement.
- Everything in Pro
- Private signal overlays for your own policy
- Committed-volume pricing
- Deployment inside your own network boundary
Why the paid tier is not a paywall
The free tier scores a payment against your own account’s history: your caps, your agents, your merchants. Everything that can be computed from what you already have is free, permanently.
The paid tier computes something different in kind: how many unrelated agents paid this merchant this month, what fraction of those payments were denied, whether five accounts that have never heard of each other all paid it inside the same five minutes, and whether a confirmed claim already exists against it. None of that is in your ledger, however large your ledger gets. It is not the same query behind a flag; it is a query over data a single customer does not have.
How metering works
One risk check is one unit, whether it arrives through /api/v1/risk-checkor through an MCP tool call. Pro includes 50,000 units a month. Above that, usage is metered at $0.01 a unit and never blocked: a hard stop in front of your agents’ payments is worse for you than an overage line is.
Stateless checks, where you send your own limits and history, are metered the same way when authenticated, and are available without an account at a lower daily cap. They are real checks, not previews: the same engine, the same ruleset hash.
On the liability guarantee
Chargeback-assist ships today: raise a claim against any scored payment and get the evidence packet. An underwritten liability backstop (Spend7 absorbing a share of a loss on a payment it scored as low risk) is not offered and is not priced. It needs an insurance and legal decision before anyone should rely on it.