Connect your portal, agents, processors, and authorization APIs to Auctra.
This guide explains how developers integrate with Auctra from account setup through real-time authorization decisions.
Integration surface
Portal plus API workflow
Portal
Accounts, teams, credentials, sandbox, production access
Enterprise APIs
Agents, spending authorities, policies, bindings, intents
Processor APIs
Real-time authorization decisions
Dashboards
Metrics, authorizations, audit trail, billing
Start here
Integration checklist
Use this sequence when connecting a new enterprise, processor, or integration environment.
Create your portal account
Sign up, verify your email, and create either an Enterprise account or a Processor account from the dashboard.
Create API credentials
Open Settings > API Credentials and generate an API key plus shared secret. Store the secret in your backend only.
Model your spend controls
Enterprises create agents, spending authorities, policies, and bindings before sending live authorization traffic.
Call the authorization APIs
Enterprises can validate an agent intent before the agent acts. Processors call the real-time authorization endpoint when a payment event arrives.
Move from sandbox to production
Test in sandbox first, then request Go-Live access from the portal. Auctra admins approve production access before live use.
Core model
Objects developers need to understand
Auctra separates who owns the spend, who initiates it, what credential is used, and which policy decides the outcome.
Enterprise
The company or platform that owns agents and defines spend controls.
Processor
The issuer, issuer processor, or payment infrastructure provider calling Auctra at authorization time.
Agent
An AI agent, workflow, service account, or automated buyer that may initiate spend.
Spending Authority
The spend-control container attached to an agent. It holds mode, status, policy, and bindings.
Policy
A template-backed rule set such as max amount, merchant lock, MCC allowlist, or human review.
Binding
The link between a payment credential, token, card, account, or instrument and a spending authority.
Intent
A preliminary validation record for an agent’s planned spend, used to check limits and policy eligibility before the agent proceeds.
Observe / Enforce
Observe approves but records what would have happened. Enforce applies the decision.
Portal setup
How teams connect through the Auctra portal
The portal is where teams create accounts, manage credentials, configure controls, review decisions, and request production access.
Enterprise portal
Enterprise users manage agents, spending authorities, bindings, policies, authorizations, activity logs, team members, and API credentials.
Open enterprise dashboardProcessor portal
Processor users review authorization traffic, inspect performance, manage credentials, and debug real-time decision requests.
Open processor dashboardProduction access
New organizations start in sandbox. Owners can request Go-Live access from the dashboard, then switch between sandbox and production after approval.
Request accessAuthentication
API requests use signed JWT bearer tokens
Create a short-lived JWT in your backend. The token's key_id must be your raw Auctra API key, and the token must be signed with your shared secret using HS256.
Get credentials
Generate API credentials from Settings > API Credentials in the portal.
Sign server-side
Never sign tokens in the browser or mobile app. Keep the shared secret in a backend secret manager.
Use bearer auth
Send Authorization: Bearer <token> on each API request.
Separate roles
Use enterprise credentials for setup and intent APIs. Use processor credentials for /v1/authorize.
import jwt from 'jsonwebtoken'
const apiKey = process.env.AUCTRA_API_KEY
const sharedSecret = process.env.AUCTRA_SHARED_SECRET
const now = Math.floor(Date.now() / 1000)
const token = jwt.sign(
{
key_id: apiKey,
iat: now,
exp: now + 3600,
},
sharedSecret,
{ algorithm: 'HS256' }
)
const headers = {
Authorization: `Bearer ${token}`,
'Content-Type': 'application/json',
}Integration paths
Enterprise and processor responsibilities
Most integrations have two sides: the enterprise side defines permissions, and the processor side asks Auctra for the live decision.
Enterprise integration
- Create or select your enterprise in the portal.
- Create one agent per automated actor or spending workflow.
- Create a spending authority for each permission boundary.
- Choose a policy template and configure parameters.
- Create one or more bindings for the card, token, account, or payment credential.
- Optionally call /v1/authorize-intent before the agent proceeds, so the enterprise can validate limits, policy fit, and purchase eligibility.
Processor integration
- Create or select your processor account in the portal.
- Generate processor API credentials with authorization scope.
- When a network authorization or payment event arrives, map it to a binding_id, instrument_id, or intent_reference.
- Call /v1/authorize with amount, currency, MCC, merchant, issuing platform, and lookup reference.
- Use APPROVE or DECLINE from the response in your payment authorization path.
- Send request IDs and correlation IDs to support when debugging production traffic.
API surface
Core endpoints
These are the primary endpoints most integrations use. The API reference contains the full list of management, dashboard, admin, and audit endpoints.
/v1/policy-templatesPublic
List available policy templates and their parameter schemas.
/v1/enterprises/{enterprise_id}/agentsEnterprise JWT
Create an agent controlled by the enterprise.
/v1/spending-authoritiesEnterprise JWT
Create a spending authority with a template-backed policy.
/v1/bindingsEnterprise JWT
Attach a payment credential or instrument to a spending authority.
/v1/authorize-intentEnterprise JWT
Run enterprise-side preliminary validation for an agent’s planned spend.
/v1/authorizeProcessor JWT
Return the real-time authorization decision for a payment event.
Authorization flows
Pre-authorization validation and real-time decisions
Use intent authorization when an enterprise wants to run preliminary checks before an agent attempts a purchase. This verifies whether the agent is allowed to proceed, whether limits remain available, and whether the planned spend qualifies under policy.
Validate an agent intent
Called by the enterprise before the agent acts. Auctra evaluates the requested amount, merchant, MCC, and spending authority policy so the enterprise can decide whether the agent should continue.
curl -X POST https://api.auctra.io/v1/authorize-intent \
-H "Authorization: Bearer $AUCTRA_ENTERPRISE_JWT" \
-H "Content-Type: application/json" \
-d '{
"spending_authority_id": "sa_123",
"amount": 4999,
"currency": "USD",
"mcc": "5734",
"merchant": "Software Store",
"intent_ttl_seconds": 86400
}'Authorize the payment
Called by the processor when the payment authorization arrives. The response contains the final decision and reason code.
curl -X POST https://api.auctra.io/v1/authorize \
-H "Authorization: Bearer $AUCTRA_PROCESSOR_JWT" \
-H "Content-Type: application/json" \
-d '{
"intent_reference": "intent_abc123",
"binding_id": "bind_123",
"issuing_platform": "MARQETA",
"amount": 4999,
"currency": "USD",
"mcc": "5734",
"merchant": "Software Store"
}'Decisions
Handling responses and reason codes
Authorization responses are designed to be simple enough for real-time payment paths and detailed enough for debugging.
Decision contract
Treat decision as the payment action. Store request_id, policy_id, binding_id, intent_reference, and reason_code for reconciliation and support.
POLICY_MATCHThe transaction matched active policy constraints.
POLICY_VIOLATION_AMOUNTThe amount exceeded the configured threshold.
POLICY_VIOLATION_MCCThe MCC did not satisfy the policy.
POLICY_VIOLATION_MERCHANTThe merchant did not satisfy the policy.
OBSERVE_ONLYThe policy would have been evaluated, but the authority is in observe mode.
NO_POLICYNo active policy was found for the referenced authority or binding.
Operational guidance
Production readiness checklist
Before sending live traffic, make sure your team has the right safeguards, logging, and environment separation.
Keep shared secrets in server-side secret storage.
Use short JWT expirations and regenerate tokens automatically.
Test observe mode before switching authorities to enforce mode.
Store Auctra request IDs and correlation IDs in your logs.
Design a clear fallback policy for network errors.
Request production access only after sandbox traffic is validated.