Hypercode is Applicable's own low-code back-end accelerator. It generates the server-side part of a web or mobile application, the database, business logic, notifications and APIs, so the part of a project that sits below the water line takes a fraction of the usual time. Clients get software that can be deployed anywhere, is fully owned by them and carries zero licence fees.
The problem: the back-end eats the budget
Every web app or mobile app has a back-end. It is the server-side software, usually hosted online, that holds the database, the business logic, the notification sending and the API endpoints the user interface talks to. It is the part below the water line, and it often consumes a large share of a project's budget. It is also data-centric, predictable and, frankly, boring to write by hand.
Applicable has been approached for years by people with good ideas: a better customer experience, a process that could be far more efficient, a new software product in their field. Too often the idea stalled on the cost of building it. We have said for a long time that our biggest competitor is people doing nothing, and doing nothing takes nobody anywhere.
What we found in the market
Through 2022 the Applicable team, led by Luna Tang and Alex Chang, evaluated no-code and low-code tools and code generators from high-end to DIY, open source to proprietary. Everything tested had at least one of the same problems: too limited, locked the client into a vendor, built on a poor choice of technology or came with high licence fees. Some tools suited simple DIY projects. None suited a development team that wanted to build real applications without those limits.
Proof of concept
By the start of 2023 Nick Sheffield had worked out an elegant way for Applicable to build its own generator, and during that summer he built the first proof of concept. Two decisions were baked in from the start.
First, Hypercode would only generate the server side. The front-end, the part in the browser or on the phone where the user experience lives, stays custom, because that is where a product is won or lost.
Second, it would not reinvent anything. Hypercode was designed to work with the modern tools developers already use. That kept the build small and means the generated code fits how a developer expects to work rather than fighting it.
Nick continued building through 2023, with Alex Chang adding the Hypercode Admin area. It was largely one person working quietly on an idea, which is how the early stage of most inventions actually goes.
What about AI?
2023 was also the year generative AI became commonplace, and the team tested ChatGPT and GitHub Copilot from the start to see whether they made a code generator redundant. They did not. AI tools can write sections of code, but their output changes from one run to the next. Once an application exists and is being iterated, the generated code has to be the same every time. The process needs to be deterministic.
AI did earn a place inside Hypercode for two jobs: suggesting new data models and letting a user ask plain-English questions of a database and get plain-English answers back.
The first projects
The first Hypercode project started in November 2023 and was delivered in January 2024, on time, on budget and with client feedback that it achieved just what they envisaged. It came in at around 50 percent of what Applicable would have estimated before the tool existed. By mid-2024 a second, larger project was under way and a third was about to start, with each one more sophisticated than the last.
What it means for clients
Applicable's Hypercode page states the client-side benefits plainly, and they are the reason it exists.
- Deployed anywhere. The generated back-end is standard code on standard infrastructure. It runs on whichever hosting suits the client, not on a vendor's platform.
- Fully owned. The client owns the code, the same as any other Applicable build. There is nothing to migrate away from later.
- Zero licence fees. Unlike the low-code platforms we evaluated, there is no per-user or per-year charge for the tool that built the software.
- Faster. The predictable part of the build takes a fraction of the time, so the budget goes into the front-end and the features people actually see.
Where it fits in an Applicable project
Hypercode does not change the process. Applicable still listens, scopes, builds and supports, and the fixed scope and price are agreed before the build starts. What changes is what that price buys. A customer portal, an internal workflow tool or a web app that replaces a spreadsheet costs less to build on Hypercode than it would have three years ago, and the client walks away owning everything. If you have an idea that stalled on cost, book a scoping call and find out what it would take now.
Key takeaways
- Hypercode is Applicable's own low-code back-end accelerator, built in-house after a 2022 review found every available tool too limited, vendor-locked or expensive to licence.
- It generates the server side only: database, business logic, notifications and APIs. The front-end stays custom.
- Generated code is deterministic, which is why generative AI tools did not replace it; AI is used inside Hypercode for data-model suggestions and plain-English database queries.
- The first Hypercode project shipped on time and on budget in January 2024 at around 50 percent of Applicable's pre-Hypercode estimate.
- Clients get a back-end that is deployed anywhere, fully owned and free of licence fees, on Applicable's usual fixed-scope process.
Faq
What is Hypercode?
Hypercode is Applicable's own low-code platform for generating the back-end of web and mobile applications: the database, business logic, notification sending and API endpoints. It combines traditional code generation with AI assistance and was built in-house in Auckland.
Is Hypercode a low-code platform I have to pay for?
No. Hypercode is a tool Applicable uses to build your software. The generated code is standard, runs on any hosting and is owned by you. There are no licence fees for the tool.
Does Hypercode build the whole application?
No. It generates the server side. The front-end, which is what users see and touch in the browser or on a phone, is still custom-built, because that is where the user experience is created.
Does Hypercode use AI?
Yes, in two places: suggesting new data models and answering plain-English questions about a database in plain English. The code generation itself is deterministic, so the same input produces the same code every time.
How much faster is a Hypercode build?
The first Hypercode project was delivered on time and on budget at around 50 percent of the cost Applicable would have estimated before the tool existed. Every project is still scoped and priced individually.
Who owns the code in a Hypercode project?
The client. As with every Applicable build, the code and the hosting accounts are yours, and the back-end can be deployed anywhere.
