Anatomy of an integration

Field Note 02 · canvas → compiled java

live execution·payment-authorization
payment-authorization · canvas
visual workflow
1
HTTP Inbound · /payments/authorize
2
Mapping · request
AuthorizePayment
function · AI risk agent
Approved 200
Declined 402
3
4
Mapping · ledger
PostToLedger
function · ISO 8583 · on-prem
5
6
Mapping · response → 200
click deploy · the canvas compiles to the java below
PaymentAuthorizationFlow.java· generated · compiled
native java · not interpreted
1package com.alphanetix.gen.payments;
2// generated from canvas · do not edit by hand
3@Compiled @Flow("payment-authorization")
4public Response authorize(HttpRequest req) {
51 var ctx = inbound.accept(req); // receive request
62 var input = map.toAuthInput(ctx.body()); // map request
73 var auth = authorizePayment(input); // 200 / 402
84 var entry = map.toLedgerEntry(auth); // map → ledger
95 var posted = postToLedger(entry, ISO_8583); // on-prem
106 return Response.ok(map.toJson(posted)); // map response → 200
11}
12private AuthResult authorizePayment(AuthInput in) { }
Runtime trace
inbound.accept() map.toAuthInput() authorizePayment() map.toLedgerEntry() postToLedger() Response.ok() · 200
1HTTP inbound
2Map request
3AuthorizePayment
4Map → ledger
5PostToLedger
6Map response → 200
One canvas → one native Java microservice. What an analyst drags, a senior engineer would hand-code in 6–12 months.
<100 ms p9910k req/sec/coreaudit per node