Know who your agent is paying. · Payment identity intelligence for AI agents

Know where the payment actually goes.

PayeeProof checks whether a machine payment endpoint, claimed service, requested resource, and public payment destination agree before an autonomous agent sends money.

Check the machine workflowLive capabilities
$0.01 per completed checkStablecoin machine rails onlyNo account required
CALLING AGENTexpected serviceexpected endpoint PUBLIC ENDPOINTHTTP 402machine metadata DESTINATIONnetwork + assetpublic pay-to PAYMENT PATH MATCHEDidentity + endpoint + destination
Identity before settlementCompare the service the agent expects with the public machine endpoint it actually reached.
Destination consistencyNormalize protocol, network, asset, resource binding, and public pay-to facts before target payment.
Credentials stay with the callerPayeeProof never needs or forwards the target merchant's payment credential, cookie, or private key.
One narrow operation

A one-cent identity check at the payment boundary.

PayeeProof is intentionally specialized. It does not score a merchant or decide whether an agent should buy. It checks whether the public machine-payment facts agree.

Paid operation
$0.01 USD
payeeproof.check_payee

Exactly one completed authorized consistency assessment. Unsafe, unsupported, or insufficient requests are rejected before PayeeProof issues its own payment challenge.

$0.01 per checkMPP / Tempox402 over HTTPstablecoin only
ID

Service match

Compare expected identity with public service and discovery metadata.

URL

Endpoint match

Bind the assessment to the exact public HTTPS target and requested resource.

402

Payment facts

Normalize the observed machine-payment challenge without moving target funds.

TO

Destination match

Surface concrete conflicts in network, asset, amount, or public payment destination.

How it works

Resolve the route before money moves.

The calling agent keeps control of the spending decision. PayeeProof provides a deterministic consistency layer immediately before autonomous machine payment.

01

Observe the target

For GET or HEAD, PayeeProof can make one bounded anonymous public probe. State-changing methods require caller-observed 402 evidence.

02

Match the payment path

Compare claimed identity, canonical resource, protocol, network, asset, amount, and public pay-to destination against supplied expectations.

03

Return bounded facts

After PayeeProof settlement, receive exact checks, provenance, continuity, limitations, next-step signals, and a deterministic SHA-256 digest.

Result classes

Signal conflicts clearly. Do not invent certainty.

Missing evidence stays unknown. A concrete contradiction is explicit. A consistent result means the available public facts agree, not that the target deserves trust.

consistent

Available public machine-readable facts agree with the expected payment identity.

review_required

Evidence is incomplete, changed, redirected, or contains a warning the caller must handle under its own policy.

conflict

Concrete observed facts contradict the expected or bound payment identity.

Machine commerce native

Built to be discovered and called by agents.

PayeeProof publishes MCP, OpenAPI, capability, well-known, skill, llms.txt, and task-guide surfaces around one stable contract. HTTP supports x402 when configured; MCP uses MPP/Tempo. There is no ordinary card checkout for the one-cent operation.

MCPOpenAPICapabilitiesllms.txtServer manifestAgent skill
Operationpayeeproof.check_payee
HTTPPOST /v1/payee/check
MCP/mcp
Price$0.01 USD
HTTP railsMPP/Tempo + x402 when configured
MCP railMPP/Tempo stablecoin
Accountnot required
Trust boundary

Payment identity intelligence, not a trust score.

PayeeProof stays useful by keeping its claims narrow and its network behavior constrained.

What PayeeProof establishes

  • public endpoint and service-identity consistency
  • payment protocol, network, asset, amount, destination, and resource-binding consistency
  • historical continuity of privacy-minimized public payment facts where available
  • explicit warnings, conflicts, limitations, and provenance

What it never claims

  • authorization to spend or a purchase recommendation
  • KYC, AML, sanctions screening, reputation, or legal identity verification
  • fraud certification, merchant honesty, delivery, or refund guarantees
  • permission to weaken SSRF controls or forward target credentials

Check the destination before the agent pays.

Start with the task guide, then use the live capability contract for machine integration.

Read task guideOpenAPI