hanteraˍ
Book a call
Platform · Integrations

The hub your order journey actually runs through.

A refund waiting on a payment provider, a delivery exception sitting in a carrier feed, a claim that takes three weeks and five systems to settle. Hantera_ receives that work, keeps the process and its state in one place, and sends the resulting action back out — so nobody has to reconstruct what happened from logs and inboxes.

Book a call → Read the ERP benchmark

Work arrives

HTTP & webhooks
Queues
Event streams
Hantera_ order & case state
rules · jobs · transformation
trace across every hop

Action leaves

Partner APIs
Queues
Event streams
One connected reality

Every system did its job. The customer is still waiting.

This is what a broken promise usually looks like from the inside. No integration threw an error. No system is obviously at fault. The refund is simply somewhere between five of them, and the only person who can see the whole thing is the customer asking about it.

  1. Mon 09:14 Return received at the warehouse event accepted
  2. Mon 09:15 Refund approved sent to the payment provider
  3. Tue 11:02 Provider is still processing no callback yet
  4. Wed 16:40 ERP posting pending waiting on the same callback
  5. Thu 08:30 “You have had my jacket for four days.” where is my money?

The agent needs one answer: what is it waiting on?

Not a log search. Not a message to finance. Not “I will find out and call you back.” The process crossed four systems and two days asynchronously — so the operation needs to see that path as one thing, including the hop that has not come back yet.

That is the difference between an integration that delivered a message and an integration that finished the job.

Delivered is not the same as done.

Hub, not just a bus

Moving the message is the easy half.

Most teams arrive here from one of two places: a web of point-to-point integrations nobody wants to touch, or middleware that routes reliably but has no idea what an order is. Both move data correctly. Neither can tell you whether the return was actually settled.

What a routing layer gives you

  • The message reached the other system
  • Retries when a delivery fails
  • A technical audit trail per connection
  • Mappings maintained away from the business

Useful — and still not an answer to “where is my refund?”

What the operation needs on top

  • The order, claim or case the message belongs to
  • State that survives days of waiting on other people
  • One trace across every asynchronous hop
  • The next action, scheduled and accounted for

The part your support and finance teams live in.

Hantera_ is the hub where both halves meet. You keep the systems you have — ERP, storefront, warehouse, carriers, payment providers — and the connective work stops being a set of scripts that only one developer understands.

What the hub is made of

Six things that decide whether a process survives contact with reality.

Real integration work is partial data, systems that answer late, peaks that arrive at the worst moment, and processes measured in weeks. These are the primitives Hantera_ gives that work — each links to the technical detail for whoever has to build it.

/ 01

Receive work however it arrives

HTTP endpoints and webhooks for request/response, queues for work that must be handled once, and retained event streams you can replay. One model, three transports.

Ingresses →
/ 02

Send action back out, safely

Outbound calls run through a typed, permission-gated egress instead of credentials scattered across workflows. One place to find and change when a partner moves an endpoint.

Egresses →
/ 03

Follow the process across the gap

When work leaves over a broker and comes back hours later, the trace travels with it — the receiving side names the sender as its parent. The slow dependency stops being a guess.

Connections & trace →
/ 04

Hold state for weeks, not milliseconds

A claim, warranty case or B2B approval outlives any single call. Ticket actors carry its state, items, relationships and history until the case is genuinely finished.

Ticket actors →
/ 05

Schedule work and absorb the peak

Move work to a quieter hour, cap how much runs at once, and watch queue depth — instead of learning about a backlog from Monday morning's contact volume.

Jobs →
/ 06

Take data in the shape it comes

JSON maps straight into a component. XML, CSV and binary payloads stream in and are parsed where they land — so a partner's format stays the partner's problem.

Parameters & parsing →

And it ships as something you can install.

An integration is rarely just a connection. It is the connection plus the rules, the scheduled work, the transformation and the screens around it. Package that as an app — versioned and upgraded on its own, without forking the platform to fit one market's paperwork.

How apps work →
Where it earns its place

Three processes that cross systems — and days.

Not integration patterns. The three situations that generate the most avoidable customer contacts in commerce operations, and what changes when the hub holds the process rather than just forwarding it.

The refund that goes quiet

  1. Return received
  2. Eligibility settled against the original order
  3. Refund sent to the payment provider
  4. Posting reconciled in the ERP
  5. Support can answer

The return event arrives on a queue and is processed once, even if the warehouse sends it twice. The refund state stays on the order instead of in a spreadsheet. A scheduled check chases the outstanding callback, and the trace shows whether the hold-up is the provider or the posting — before the customer asks a second time.

The delivery that goes wrong

  1. Carrier events stream in
  2. Matched to the delivery and the promise
  3. Exception recognised
  4. Recovery decided
  5. Customer told first

Carrier feeds are high volume and occasionally malformed. A stream ingress keeps its own durable position, so a restart resumes where it left off rather than replaying a week of events — and one unparseable event is quarantined instead of stalling the feed behind it. The brand gets to tell the customer before the tracking page does.

The claim that takes three weeks

  1. Claim raised
  2. Evidence and inspection gathered
  3. Approval from the right people
  4. Replacement or credit issued
  5. Settled and reconciled

Warranty cases, B2B disputes and recalls stall because nothing owns the middle. A ticket holds the case state, the linked order, the documents and the decisions, while jobs handle deadlines and follow-ups. The process survives handovers, holidays and the person who normally deals with it being away.

Across the stack

Keep your systems. Connect the promise.

Hantera_ sits between the systems you already run rather than replacing them — so becoming the hub is an integration project, not a replatforming one.

ERP & finance

Keep financial, tax, payment and order records connected to the operational decisions that change them.

Ecommerce & POS

Online, store, inside sales and marketplace channels working from shared order context.

Warehouse & logistics

Routing, shipments, carriers and delivery events attached to the order journey they affect.

Help desk & returns

Order and recovery context where support already works — Zendesk, CustomerFirst and more.

Returns & Recovery →

Tax & payments

Certified AvaTax integration, with captures and refunds driven from the order itself.

Payments & Financial Resolution →

Marketplaces & partners

External channels, drop-ship partners and franchise systems held to the same operational reality.

Hani, the Hantera_ mascot, at a switchboard patching several commerce systems into a single order line.

Proof, not promises.

Didriksons replaced their ERP and their commerce stack in the same quarter, with Hantera_ as the hub every other system met in. The Business Central integration itself took 37.25 hours — against an industry benchmark of 200–600 for equivalent scope.

Read the case study →

Bring the process that keeps falling between systems.

Show us one refund, delivery exception or claim that takes too long to settle. We'll map the events, the state and the systems involved — and tell you what we would connect, what we would leave alone, and roughly what it takes.

Book a call → Read the integration docs