Most businesses run on several tools that do not talk to each other. Someone copies orders from the online store into the accounting software, or retypes new leads from a form into the CRM. API development fixes that. We build the connections that move data between systems automatically, and we build APIs for your own product when customers or partners need to connect to it.
Most business software sold today has an API. The work is in connecting those tools so the data stays correct when something goes wrong.
Two kinds of API work
Building your own API. Your app needs a way for its own screens, a mobile app or partner companies to read and change data. We design the API, write it in Node.js and TypeScript, and document it. Most of our APIs follow the REST style, which is the common pattern where each type of data has its own web address, such as /orders or /customers.
Third party API integration. You want tools made by other companies to share data. For example, you might want new Shopify orders to appear in QuickBooks Online, or paid Stripe invoices to update a customer record in HubSpot. We write the code that listens for changes in one tool and sends them to the other.
Many projects need both. A customer portal might have its own API for its screens and also pull invoices from your accounting software.
Tools we connect
These are the kinds of systems businesses most often ask us to link up:
- CRM. HubSpot, Salesforce, Pipedrive and similar tools for contacts and sales.
- Accounting. QuickBooks Online, Xero and FreshBooks for invoices, payments and tax codes such as HST and GST.
- E-commerce. Shopify and WooCommerce orders, stock levels and customers.
- Booking and scheduling. Calendars, appointment tools and staff rosters.
- Payments. Stripe for cards, subscriptions and refunds.
- Email and messaging. Receipts, alerts and follow-ups sent when something happens in another tool.
If a tool you use is not on this list, ask. Most business software publishes its API documentation, and we read it before we quote.
Keeping data in sync
Syncing data sounds simple until two systems disagree. We settle these questions before we write code:
- Which system is the source of truth for each type of record, such as customers, prices or stock.
- What happens when the same record changes in both places at once.
- How to match records that have no shared ID, for example by email address.
- How to load the old records that existed before the integration started.
We also use webhooks where a tool offers them. A webhook is a message a tool sends the moment something happens, such as “invoice paid”. It is faster than checking for changes every few minutes and uses fewer requests.
Controlling who can reach your data
An API is a door into your data, so we treat access carefully.
- Each app or partner gets its own key or login, with only the permissions it needs.
- For tools that let users connect their own accounts, we use OAuth. OAuth is the standard “Connect your account” screen, which gives access without anyone sharing a password.
- Keys and secrets are stored in the hosting platform’s secret storage, never in the code.
- Rate limits cap how many requests a caller can make per minute. This protects your API from a buggy script or a misuse of access.
Other companies’ APIs have rate limits too. When we sync a large number of records, we spread the requests out so your account does not get blocked.
Documentation other developers can use
If partners or other developers will use your API, the documentation matters as much as the code. We write an OpenAPI description, which is a standard file that tools can turn into readable reference pages and code samples. We add plain examples for the common tasks and a test environment with sample data, so a partner can try requests without touching real records.
When the API needs to change later, we add a new version and keep the old one running for a set period. That way a change on your side does not break someone else’s app overnight.
When an integration breaks
Integrations fail for ordinary reasons. A login token expires. A provider retires an old version of its API. Someone renames a field in the CRM. The worst failures are silent ones, where data stops moving and nobody notices for weeks.
We build every integration to fail loudly and recover on its own where it can:
- Failed requests are retried automatically, with a growing wait between tries.
- Records that still fail are kept in a queue so none are lost, and they can be replayed once the cause is fixed.
- Errors send an alert to you or to us by email or Slack.
- Every sync is logged, so we can see exactly what was sent and when.
If you already have an integration that drops orders or creates duplicates, we can take it over. We start by reading the logs and the code, then fix the cause of the problem.
Pricing and next steps
A new API or a set of connected systems is scoped as a custom web app project. Prices are on our pricing page.
APIs are often the first step toward other work. You might want AI integration that reads data from your systems, or an AI agent that acts on it. A dashboard can show the combined data in one place. Our Node.js development page covers the backend work in more detail.
Tell us which tools you use and where data gets stuck. Contact us and we will reply with the questions we need answered before we quote.
