An API connector is the software layer that turns an API into a working connection between two systems, handling the authentication, data mapping and back-and-forth an API leaves to you. Most business platforms expose an API, yet having one does not mean your systems are connected. This guide explains what a connector is, how it differs from an API and how businesses use one to link tools like a CRM and accounting system.
Table of Contents
What is an API connector?
An API connector is a reusable component that sits on top of an API and turns it into a usable connection between systems. Where an API defines how a system can be accessed, the connector manages the authentication, the requests and the data handling, so a workflow or app can exchange data without a developer wiring every call by hand. Our guide to what an application interface is covers the API side in more detail.
Microsoft describes a custom connector as a wrapper around a REST or SOAP API that lets platforms like Azure Logic Apps and Power Automate communicate with that API. The useful part of that model is what the connector exposes: triggers, which are events that start a workflow such as a new deal being won, and actions, which are tasks such as creating an invoice. Once a system is wrapped in a connector, non-developers can build integrations by joining those triggers and actions together, rather than writing and maintaining raw API calls.
What is the difference between an API and an API connector?
An API is the interface a system exposes so other software can request its data or services. An API connector is the built layer that uses that API to move data reliably between systems. The API is the doorway. The connector is what walks through it on a schedule, handles the login and keeps the data mapped correctly.
IBM defines API integration as using APIs to connect enterprise applications, systems and workflows for the exchange of data and services. The gap between having an API and being integrated is where most businesses get stuck, and the numbers show how wide it is. MuleSoft’s 2026 Connectivity Benchmark Report found the average organisation now runs 957 applications and only 27% of them are connected. That gap between owning software and having it work together is the whole reason business systems integration exists as a discipline, and connectors are the practical tool that closes it.
How do API connectors work behind the scenes?
An API connector works by creating an authenticated connection to a system, then using that connection to run triggers and actions against its API. It authenticates once, maps the fields between the two systems, then either waits for an event or runs on a schedule and passes the data across, retrying and logging so nothing is lost quietly.
The first step is the connection. Most connectors require you to create a connection and authenticate your identity before they will do anything, usually through OAuth or an API key held in a secure store rather than pasted into code. From there the connector listens for its trigger, then runs its actions, and the real work happens in the middle: mapping one system’s fields to the other’s. That mapping layer is where integrations are made or broken, because two systems almost never describe the same customer or invoice the same way. In the HubSpot to MYOB build below, that meant mapping a HubSpot deal to MYOB line items, tax codes and contact records before a single record moved. The connector itself usually runs inside middleware such as Power Automate, Azure Logic Apps or an Azure Function, which is also where retries, error handling and logging live. We cover the same pattern for Microsoft’s document platform in our SharePoint API guide.
What are the main types of API connectors?
API connectors come in two main types, prebuilt and custom. Prebuilt connectors are ready-made for popular services and work with little or no code. Custom connectors are ones you build when no prebuilt option exists for the system or the scenario you need. Most real integrations end up using a mix of the two.
Microsoft’s connector documentation lists exactly these two types, prebuilt and custom, and, per the Azure Logic Apps documentation, the platform provides more than 1,400 prebuilt connectors for common services. A prebuilt connector is fast to use and maintained by the platform vendor, so it is the default choice whenever one exists for the system you are connecting. A custom connector is what you reach for when the system is niche or in-house, or when a standard connector cannot handle the specific fields or logic your process needs. The table below sets the two side by side on the factors that decide the choice in real projects.
| Factor | Prebuilt connector | Custom connector |
|---|---|---|
| What it is | A ready-made connector for a popular service | A connector you build over an API with no prebuilt option |
| Setup | Create a connection and authenticate, little or no code | Describe the API, define the actions, then secure it |
| Best for | Common tools like Office 365, Salesforce and mainstream CRMs | Niche or in-house systems, or logic a standard connector cannot handle |
| Maintenance | Maintained by the platform vendor | You maintain it as the underlying API changes |
| Control | Limited to the actions the vendor exposes | Full control over actions, fields and error handling |
When do you need a custom API connector?
You need a custom API connector when no prebuilt connector exists for a system, or when the prebuilt one cannot handle the specific fields, volume or logic your integration requires. It is also the right call when two systems both have an API but nothing off the shelf makes them talk in the way your actual process works.
This is more common than the marketing around no-code tools suggests. MuleSoft’s 2026 report found 82% of IT leaders cite data integration as one of the biggest challenges they face in using AI, and teams still report spending an average of 36% of their time designing, building and testing custom integrations. The reason custom work persists is simple: real business processes rarely match the generic path a standard connector assumes. When the mismatch is in the data itself, how customers, invoices or jobs are identified across two systems, a prebuilt connector often gets you eighty per cent of the way and leaves the awkward twenty per cent that actually matters.
What good API connectivity looks like: connecting HubSpot to MYOB
Good API connectivity is invisible when it works. For a Sydney construction firm, a won deal in HubSpot became a customer and invoice in MYOB automatically, with no one re-keying anything. We built it on Azure Logic Apps, and the hard part was not the connection, it was matching records and mapping fields between two very different systems.
The problem was the kind that hides in plain sight. Sales ran in HubSpot and accounts ran in MYOB, while the office manager re-keyed every won job from one into the other by hand. It was slow, and worse, it created mismatched and duplicated records whenever a name was typed slightly differently. Both systems had a capable API, so the raw connectivity was never in doubt. What the business needed was a connector that understood its data. We built that as an Azure Logic Apps integration, the kind of CRM integration work that turns two separate tools into one process.
The workflow runs in six steps:
- A won deal in HubSpot triggers the Logic Apps workflow automatically.
- The workflow checks MYOB for a customer that matches the HubSpot company, using ABN and name, before it creates anything.
- If no match exists it creates the customer, and if a match exists it updates the existing one, so no duplicates are made.
- It maps the HubSpot deal to a MYOB invoice, including line items, tax codes and the contact details, so the invoice is correct on the first pass.
- It writes the MYOB invoice number back into HubSpot, so the sales team can see the accounting status without leaving their CRM.
- It logs every run and retries on transient errors, so a brief MYOB timeout never loses a deal.
The two genuinely hard parts were the ones that never show up in a demo. The first was duplicate and record matching. A construction client might already exist in MYOB from an earlier job, so the connector had to match on ABN and name and decide whether to create or update, rather than blindly adding a second record. The second was field and data mapping. HubSpot and MYOB structure a contact, an address and a tax code differently, so each field had to be mapped deliberately and tested against real records, not sample data. Get those two right and the rest is plumbing. The result was that the double entry disappeared entirely. Every won deal now flows across on its own, the office manager stopped re-keying invoices by hand and the mismatched records that used to surface at reporting time stopped appearing.
How do you choose the right API connectivity approach?
Choose your API connectivity approach by matching the tool to the volume, the complexity and who maintains it. A prebuilt connector in a no-code tool suits simple, common links. A custom connector or Azure Logic Apps suits anything with real volume, custom fields or error handling. Custom code is the last resort, for logic nothing else can express.
The decision usually comes down to five questions: is there a prebuilt connector for both systems, how much data moves and how often, how different are the two data models, how much does it matter if a record fails silently and who will own the integration once it is live. The more you answer towards volume, difference and consequence, the further you move from a simple prebuilt connector towards a managed integration. The table below is the shortlist we walk clients through.
| Approach | Best for | Watch out for |
|---|---|---|
| Prebuilt connector in a no-code tool | Common apps, simple triggers, low volume | Limited to the vendor’s actions and fields |
| Custom connector or Azure Logic Apps | Custom fields, real volume, retries and logging | Needs design and ongoing maintenance |
| Custom code with Azure Functions | Complex logic or transformation nothing else handles | The most developer effort and the highest maintenance |
Getting this call right early is most of the value, because the wrong approach is expensive to unwind once data is flowing through it. Our API and integration services work starts by scoping exactly these questions before a line of the connector is built, and it sits within our broader system integration and CRM practice.
Frequently asked questions
What is an API connector in simple terms?
An API connector is a ready-made bridge between two software systems. It uses one system’s API to send and receive data on your behalf, handling the login, the requests and the formatting so the two tools can work together. Think of the API as a doorway a system offers, and the connector as the thing that walks through it reliably, on a schedule and without a person copying data across by hand.
Is an API connector the same as an API?
No. An API is the interface a system publishes so other software can request its data or services. An API connector is the built layer that actually uses that API to move data between systems. The API defines what is possible, and the connector does the work. A system can have a capable API and still not be connected to anything, because nothing has been built to use it yet.
Do I need an API connector if my software already has an API?
Usually yes. Having an API means a system can be connected, not that it is. Something still has to authenticate, map the fields and move the data between your systems, and that something is a connector or an equivalent integration. The API is necessary but not sufficient. MuleSoft’s 2026 research found the average organisation runs 957 applications and only 27% are actually connected, which is the gap connectors fill.
What is an API connection in Azure?
In Azure, a connection is the authenticated link a connector uses to reach a service. When you add a connector such as Salesforce or SQL Server to an Azure Logic Apps workflow, you first create a connection and sign in, and that saved connection is then reused to run the connector’s triggers and actions. Azure Logic Apps offers more than 1,400 prebuilt connectors, and each one relies on a connection you authorise once.
How much does it cost to connect two systems with an API connector?
It depends on the systems and the complexity, not on a fixed connector price. A simple link between two mainstream tools with prebuilt connectors can be quick. A custom integration with record matching, field mapping and error handling is a larger build. The ongoing costs sit in the hosting, such as Azure Logic Apps or Functions billed on usage, and in maintaining the integration as the underlying APIs change. We scope each one before quoting.
Can HubSpot and MYOB be connected?
Yes. Both HubSpot and MYOB expose an API, so they can be connected so that a won deal in HubSpot becomes a customer and an invoice in MYOB automatically. The work is not the connection itself, it is matching records so you do not create duplicates and mapping fields like tax codes and contacts correctly. We built exactly this for a Sydney construction firm using Azure Logic Apps, and it removed their double entry entirely.
Turning APIs into connections you can trust
An API is only potential. The connector is what turns that potential into a connection you can rely on, one that authenticates itself, maps the data correctly and keeps two systems in step without a person in the middle. Most businesses do not have an API problem, they have a connectivity problem, and the fix is rarely more software. It is the deliberate work of connecting the software they already own.
If you are trying to connect a CRM and an accounting system, or any two tools that both have an API but still are not talking, that is the work we do every week for Australian organisations. Book a free call and we will talk through your systems and the cleanest way to connect them, with no obligation.
