Agents
Autonomous actors such as AI agents, automation systems, or software services. Identity, credentials, API authorization, delegation targets. Stable ID (e.g. agt_xxx) and signing credentials.
The trust and transaction layer for autonomous systems.
AffixIO provides infrastructure for machines, devices, and AI agents to perform authorized transactions and actions with verifiable proofs. This is not a payment gateway in the traditional sense: it is machine commerce infrastructure: the layer that enables agentic payments, autonomous transaction systems, and AI payment infrastructure at scale.
The platform enables:
Related: Agentic AI payments (ZK on the edge, offline proofs), Identify via API (binary eligibility), Merchant SDK (offline terminal payments).
In this page: What is agentic infrastructure, core primitives, V1 API endpoints, offline transaction architecture, cryptographic proofs, developer flow, FAQ, build with AffixIO.
A clear definition for developers and technical documentation.
In practice, agentic payment infrastructure and autonomous payment infrastructure are the same idea: a trust and authorization layer that sits between machines and the services they interact with. It answers: who is acting, is the action allowed, what limits apply, and how can the action be proven later.
Autonomous systems increasingly perform actions such as making purchases, requesting services, unlocking physical devices, interacting with APIs, and executing workflows. These systems require a layer that determines who the actor is, whether the action is allowed, what limits apply, and how the action can be proven later. AffixIO provides this layer.
AffixIO acts as a trust and authorization system between machines and the services they interact with. It is the infrastructure that enables AI agent payments, machine-to-machine payments, and autonomous payments without requiring a human in the loop at the point of execution.
These are the building blocks available in the platform for building agentic payment infrastructure and machine commerce applications.
Autonomous actors such as AI agents, automation systems, or software services. Identity, credentials, API authorization, delegation targets. Stable ID (e.g. agt_xxx) and signing credentials.
Terminals, kiosks, robots, IoT hardware. Device identity, attestation, pairing with agents, offline support. Operate disconnected with local proof generation and transaction queueing.
Organisations or users grant machines permission to act. Scope (currencies, limits, merchants), validity windows. Verify before an action, revoke at any time.
Evaluates whether actions are allowed. Rule evaluation, scope checks, policy enforcement, binary allow/deny/review. Consumes delegation state and device attestation.
Transaction orchestration: intent creation, trust authorization, payment execution, refund/reversal. Lifecycle: created → authorised or denied → captured or cancelled.
Public API routes available in AffixIO V1. All paths are relative to the API base (e.g. https://api.affix-io.com/v1). Authentication is via X-API-Key or Authorization: Bearer <key> unless noted.
Health check. No authentication required. Returns database connectivity status.
Create a new autonomous agent (name, type, ownerOrgId, metadata).
Retrieve agent details.
Update agent configuration (name, metadata, status).
Issue API or signing credentials for the agent. Secret returned only at creation.
Rotate credentials and revoke previous ones.
Create a device (name, type, ownerOrgId, optional agentId, capabilities, metadata).
Retrieve device details.
Update device configuration.
Record or verify attestation evidence.
Pair device to an agent or terminal identity.
Create a delegation (delegator, delegate, scope, startsAt, expiresAt).
Get delegation details.
Revoke delegation immediately.
Verify delegation for a proposed action (e.g. payment.create, amount, currency, merchantId).
Evaluate whether an action is allowed (subjectType, subjectId, action, context). Returns decision, reasons, proofId.
Create a trust policy (name, rules).
Get trust policy.
Evaluate a policy against input context.
Create payment intent (payer, payee, amount, currency, mode: online|offline). Idempotency key supported.
Get payment intent.
Run trust check and mark intent authorised or denied.
Finalise into a payment record.
Cancel payment intent.
Create a payment (validates trust internally). Idempotency key supported.
Get payment.
Refund payment.
Reverse payment.
Endpoints for offline transaction infrastructure: submit transactions, commit, sync, queue visibility, nullifier check and reconcile. See Offline transaction architecture.
Generate and verify cryptographic proofs (payment, eligibility, delegation, device attestation, compliance). Endpoints: generate, verify, get by ID, aggregate, sync, revoke; convenience endpoints for payment, eligibility, consent, device-attestation, compliance.
Short-lived machine-usable tokens (trust, eligibility, payment_authority, device_session, delegation). Generate, validate, revoke, get by ID, exchange.
Append-only audit events; filter by subjectId, type, date; attach proof references; verify integrity.
Create organisation (setup).
OpenAPI 3.0 specification. No authentication required for openapi.json.
Devices and AI agents often operate in disconnected or unreliable connectivity environments. AffixIO's offline payment infrastructure supports local proof generation, transaction queueing, nullifier-based double-spend prevention, and later synchronization with the AffixIO API.
Example environments: aircraft, events, transport, remote retail, and autonomous systems in the field. This is the same model that supports robot payments and IoT payment systems where the device cannot assume real-time connectivity. The combination of local proof generation, queueing, and nullifier-based reconciliation is what makes offline payment infrastructure for the machine economy viable: devices remain operational while disconnected, and the system remains secure against double-spend when they sync.
AffixIO supports cryptographic proofs for actions and transactions. Without going deep into mathematics: proofs allow systems to verify that a transaction occurred, that the action was authorized, and that verification can be done without storing sensitive data. Proof systems enable privacy-preserving verification: third parties (auditors, regulators, other machines) can confirm that a decision or payment was valid without seeing the underlying credentials or PII. In V1, proofs use signed payloads or HMAC-based formats; the design allows more advanced proof systems (including zero-knowledge) to be integrated later.
AffixIO acts as the trust and authorization layer between autonomous systems and the external services they use. Machines identify themselves, obtain delegated authority, and request trust decisions; AffixIO evaluates and returns allow/deny and optional proofs. Payment orchestration and proof generation sit in this layer so that downstream systems can rely on a single control plane for machine commerce infrastructure.
A simplified workflow for building on AffixIO agentic payment infrastructure:
Example using the API (conceptually):
// 1. Create org and API key (platform setup)
// POST /v1/orgs → POST /v1/api-keys
// 2. Register agent
const agent = await fetch('https://api.affix-io.com/v1/agents', {
method: 'POST',
headers: { 'Content-Type': 'application/json', 'X-API-Key': API_KEY },
body: JSON.stringify({
name: 'Checkout Agent',
type: 'workflow',
ownerOrgId: 'org_xxx',
metadata: {}
})
}).then(r => r.json());
// 3. Create delegation (allow agent to spend)
await fetch('https://api.affix-io.com/v1/delegations', {
method: 'POST',
headers: { 'Content-Type': 'application/json', 'X-API-Key': API_KEY },
body: JSON.stringify({
delegatorType: 'org',
delegatorId: 'org_xxx',
delegateType: 'agent',
delegateId: agent.id,
scope: { canMakePayments: true, allowedCurrencies: ['GBP'], maxSingleAmount: 100 },
startsAt: new Date().toISOString(),
expiresAt: new Date(Date.now() + 86400000).toISOString()
})
});
// 4. Trust check before payment
const trust = await fetch('https://api.affix-io.com/v1/trust/check', {
method: 'POST',
headers: { 'Content-Type': 'application/json', 'X-API-Key': API_KEY },
body: JSON.stringify({
subjectType: 'agent',
subjectId: agent.id,
action: 'payment.create',
context: { amount: 50, currency: 'GBP' }
})
}).then(r => r.json());
// trust.decision === 'allow' → proceed
// 5. Create payment intent and capture
const intent = await fetch('https://api.affix-io.com/v1/payment-intents', { ... }).then(r => r.json());
await fetch(`https://api.affix-io.com/v1/payment-intents/${intent.id}/authorise`, { method: 'POST', ... });
await fetch(`https://api.affix-io.com/v1/payment-intents/${intent.id}/capture`, { method: 'POST', ... });
Example builders on agentic payment infrastructure and machine commerce infrastructure:
AI systems increasingly perform autonomous tasks: purchasing services, accessing APIs, unlocking physical infrastructure, executing automated logistics, and interacting with other machines. These interactions require transaction infrastructure similar to what the internet required for human commerce. The machine economy: and future AGI economy infrastructure: depends on a trust layer that can verify identity, enforce permissions, evaluate policies, and produce verifiable proofs.
AffixIO provides the trust layer that enables machines to safely participate in economic systems. This is machine commerce protocol and autonomous financial infrastructure in practice: not a replacement for payment rails, but the control plane that authorises and proves machine-initiated transactions. As the AI agent economy and agentic commerce grow, the need for AI transaction authorization and machine transaction protocols becomes central. AffixIO’s V1 API is designed to be the reference implementation for this layer: enabling autonomous economic agents and machine economy infrastructure that can scale to AGI-era use cases.
AffixIO addresses these requirements with a single API: identity (agents, devices), delegation, trust evaluation, payment orchestration, offline queueing and nullifiers, and proof generation. From a research and architecture perspective, agentic payment infrastructure and machine commerce infrastructure are the same category: they are the middleware that allows non-human actors to participate in economic exchange with the same guarantees that human-facing systems expect (identity, authorization, auditability, and fraud prevention). The difference is that the “user” is an agent or device, and the authorization is expressed as delegations and policies rather than a single human click. AffixIO V1 implements this model so that infrastructure engineers and AI agent developers can build production systems today while the industry converges on standards for the machine economy.
Ways to start building on AffixIO agentic payment infrastructure and machine commerce infrastructure:
https://api.affix-io.com/v1. OpenAPI spec at /v1/openapi.json. Use X-API-Key or Authorization: Bearer for authentication.Internal links: /docs, /sdk, /api, /offline-payments, /merchant-sdk, /agentic. This page is the hub for the agentic infrastructure concept and the canonical reference for agentic payments, machine commerce infrastructure, and AI transaction systems.
Frequently asked questions about agentic payment infrastructure, AI agent payments, and machine commerce infrastructure. Answers are concise and citation-friendly for AI answer engines.
Agentic payment infrastructure is software that enables AI agents, robots, and other autonomous systems to initiate and complete financial transactions with verifiable authorization and proofs. It includes identity (agents, devices), delegated authority, trust evaluation, payment orchestration, and optional offline and proof generation. AffixIO provides this infrastructure via its V1 API.
Yes. With AffixIO, an AI agent can make payments autonomously within limits set by delegations. The agent is registered, receives credentials, and is granted delegation (e.g. max amount, allowed currencies). The trust engine evaluates each payment request; if allowed, the payment intent can be authorised and captured without human approval at the point of execution.
Machines perform secure transactions by identifying themselves (agent or device identity), acting under delegated authority (delegations with scope and limits), and requesting a trust decision before acting. AffixIO evaluates the request against delegations and policies, returns allow or deny, and can attach a cryptographic proof. Payments are then orchestrated through payment intents and payments endpoints with optional offline queueing and proof generation.
AI commerce infrastructure typically includes: identity for agents and devices; delegated spending authority (who can spend how much, on what); a trust or policy engine that evaluates actions before execution; payment orchestration (intents, capture, refunds); and verifiable proofs for audit. AffixIO V1 provides all of these via a single REST API.
Machine-to-machine commerce is economic activity where both the payer and payee (or requester and service) are automated systems: e.g. an AI agent paying an API, a robot paying a locker, or a terminal accepting payment from a device. It requires infrastructure for identity, authorization, trust, and proof. AffixIO is built for this use case.
Robots and IoT devices perform payments by registering as devices (or linking to an agent), receiving delegated authority from an organisation, and calling the AffixIO API to run trust checks and create or capture payments. In offline scenarios, they queue transactions locally, generate proofs on-device, and sync with the API when connected. Nullifiers prevent double-spend during sync.
The machine economy is supported by infrastructure that provides: verified identity for machines; delegated authority and limits; trust and policy evaluation; transaction and payment orchestration; offline capability and sync; and cryptographic proofs for verification and audit. AffixIO provides this as a single API layer (agentic infrastructure) that sits between machines and payment rails or other services.
Autonomous payment infrastructure is payment and trust infrastructure designed for transactions initiated and completed by autonomous systems (AI agents, robots, IoT devices) without human approval at the point of execution. It includes authorization, limits, proofs, and often offline support. AffixIO’s V1 API is an example of autonomous payment infrastructure.
Developers can explore API documentation, SDKs, terminal integrations, and agent frameworks to build on AffixIO agentic payment infrastructure and machine commerce infrastructure.