Systems

Bespoke systems for operations that do not fit off-the-shelf software

When the process is the company’s edge, bending it to fit a generic system means giving up the very thing that makes it competitive. We build the system around the operation — not the other way round.

When it makes sense

Not every problem calls for a new system

Let us be blunt: if there is a market product that solves your case well, buying it is cheaper, faster and less risky. We sell ready-made systems ourselves, and that is frequently what we recommend.

Bespoke development is justified when at least one of these is true:

  • The process is the competitive edge. Bending it to generic software means levelling down to the competition.
  • What the market offers solves 70% and the rest ends up in spreadsheets. Parallel spreadsheets are the classic symptom.
  • The company operates in a niche too small to attract specialised products.
  • Several systems need to talk to each other and the integration is more complex than the systems themselves.
  • The business rules are dense — tax, contractual, regulatory — and change often.
  • Per-user licensing stopped making sense as the operation grew.

If none of this applies, we say so in the first meeting.

It starts with a conversation

Before a single line of code

No project starts with a proposal. It starts with a meeting.

Its purpose is not to sell: it is to understand the operation and size the problem. We want to know how the work happens today, where it jams, who does what, which decisions depend on which information, and what has already been tried.

One of three things comes out of that conversation:

1

The recommendation to build nothing.

A market product covers it, or the problem is one of process rather than software. It happens, and we say so.

2

A reduced scope.

Often 20% of what was imagined solves 80% of the pain. Better to deliver that in two months than the rest in two years.

3

A sized project.

Scope, phases, technologies, timelines and investment, with clarity about what is in and what was left for later.

Only after that stage do we talk about a proposal. It is the conversation that avoids the worst possible outcome in bespoke development: competently building the wrong solution.

The architecture

Separated into layers, always

This is the technical decision with the biggest impact on your system’s long-term cost — and the one that shows up least in a commercial proposal.

We build every solution with independent layers: business rules on one side, data access on another, interfaces on a third. They communicate through defined contracts, not direct coupling.

One architecture, many interfaces

The layers are independent: each can change without dragging the others along.

Desktop — WinForms · WPF · Electron Desktop Integrações — ERPs · webhooks · EDI Integrações SQL Server — ambientes Microsoft SQL Server PostgreSQL — relacional · open source PostgreSQL Web — React · Vue · Angular · TypeScript Web Mobile — React Native · Flutter · MAUI Mobile API REST API REST Hibernate Hibernate Dapper Dapper EF Core EF Core CORE Business rules C# · .NET Java · Spring
  • Interfaces
  • Contracts
  • Persistence
  • Data
  1. Interfaces

    • Web React · Vue · Angular · TypeScript
    • Mobile React Native · Flutter · MAUI
    • Desktop WinForms · WPF · Electron
    • Integrações ERPs de terceiros · webhooks · EDI
  2. Contracts

    • API REST contratos entre camadas
  3. Persistence

    • EF Core domínio, migrações e escrita
    • Dapper relatórios e alto volume
    • Hibernate / JPA equivalente no ecossistema Java
  4. Data

    • PostgreSQL relacional · open source
    • SQL Server ambientes Microsoft

The technology in each layer is chosen per project, to match the client’s environment. What never changes is the separation between them.

  • the same backend serves every interface
  • swapping the database does not rewrite business rules

Why this matters in practice

One backend, many interfaces

If today you need a web system and in a year an app for the field team, the app consumes the same backend. No business rule is rewritten, no logic duplicated, and you never end up in the classic situation where web and mobile compute the same figure in slightly different ways.

Swapping the database does not rewrite the system

The data-access layer isolates the database from everything else. Migrating from SQL Server to PostgreSQL, or the reverse, is localised work.

Integrations come through the front door

A third-party ERP, a marketplace or a partner system consumes the same API your front-end does. There is no “integration hack” because there is no need to invent a parallel path.

Testing becomes possible

A business rule isolated from interface and database can be tested automatically. A rule scattered inside a screen cannot.

The technologies

What we use, and why

We do not commit to a single stack out of dogma. Each layer has options, and the choice is made per project, considering the client’s existing technology, the team that will live with the system, and the nature of the operation.

Database

PostgreSQL
Relational, open source, no licence cost. Excellent for JSON, geospatial data and analytical workloads. It is our default recommendation for new systems, especially cloud-deployed ones.
SQL Server
When the client already has Microsoft infrastructure and licensing, a team used to the tooling, or integration with legacy systems that already live there. Strong native integration with the .NET ecosystem.

The choice is usually less technical than it looks: what the company already has, and who will administer the database day to day, weigh more.

Backend

C# with .NET
A mature, performant platform with strong typing and excellent tooling. It is where we have the largest body of accumulated experience.
Java with Spring
A robust ecosystem, well established in corporate environments and highly compatible with infrastructure already in place at larger companies.

Both solve the same problems with equivalent quality. The decision weighs what the company already runs and who will maintain it in future.

Data access: ORM

An ORM (Object-Relational Mapper) is the layer that translates between the world of objects, where the code lives, and the world of tables, where the data lives. Without one, every query becomes hand-written SQL and every result becomes repetitive conversion code.

Entity Framework Core (.NET)
A full ORM. It maps entities, generates queries, manages schema migrations and tracks changes automatically. It speeds up CRUD and domain-rule development considerably.
Dapper (.NET)
A micro-ORM. You write the SQL and it handles the mapping to objects. More control, less automation, performance close to raw ADO.NET.
Hibernate / JPA (Java)
The mature equivalent in the Java ecosystem, with advanced caching and mapping features.

Frontend

Web applications
React, Vue or Angular, with TypeScript. React is our default for ecosystem maturity and hiring availability; Vue when the client team already knows it; Angular in corporate environments that have standardised on it. TypeScript always — typing on the front end eliminates an entire class of errors that would otherwise only surface in production.
Mobile apps
React Native or Flutter for code shared between Android and iOS; .NET MAUI when the backend is .NET and the team writes C#. Pure native only when the project needs specific hardware capability that justifies the cost of two codebases.
Desktop applications
Not obsolete technology — the right choice in several scenarios: operations without reliable internet, integration with hardware (scales, scanners, fiscal printers, turnstiles), heavy keyboard use by a trained operator. WinForms and WPF in the .NET world, Electron where sharing code with the web version makes sense.

The front-end choice does not lock in the others. That is the practical advantage of layered separation: starting with web and adding mobile later is natural evolution, not a restart.

Who builds it

Three engineers, three specialities that do not overlap

Bespoke projects are settled by the repertoire of whoever builds them. Our team combines deliberately different profiles.

Management-systems engineer

  • Thirty years building ERPs for furniture stores, distributors, travel agencies and companies across a wide range of industries.
  • Deep experience in system integration — including scenarios with no API and no documentation.
  • Studying Accounting, which brings a rare combination: someone who understands double-entry bookkeeping, chart of accounts, accrual basis and tax assessment and writes the code that implements it.

Most serious defects in business systems are not programming errors — they come from the programmer not understanding accounting.

Architecture and innovation engineer

  • Speaker and agile methods specialist, working on structuring development processes and delivery cadence.
  • Artificial intelligence consultant for developers, CEOs and CIOs — work that involves both technical application and strategic translation for decision makers.
  • Cryptography and blockchain specialist, with applied experience in agribusiness — a sector with its own demands around traceability, seasonality and field integration.

Systems and infrastructure engineer

  • Electronics and network equipment specialist — hardware knowledge that rarely coexists with software development, and that is decisive when the system talks to devices.
  • Developer of systems for internet service providers, with command of network management, provisioning and recurring billing.
  • Developer of electronic time-clock systems, an area combining strict labour legislation, integration with certified equipment and zero tolerance for calculation error.

How we run it

From the first meeting to live operation

  1. Discovery meeting

    No cost, no commitment. Understanding the operation, the problem and the context.

  2. Sizing

    Scope, phases, technologies and investment. What goes in now and what waits.

  3. Prototyping

    Screens and flows before code. Changing your mind is far cheaper at this stage — and this is where the client discovers what they actually wanted.

  4. Development in cycles

    Partial, working deliveries at short intervals. You watch the system grow and correct course along the way, instead of receiving everything at the end and finding out it was not what you meant.

  5. Acceptance with the people who will use it

    Testing with the real team, real data and real volume. Not with the manager who approved the project, but with whoever will open that screen fifty times a day.

  6. Rollout and support

    Data migration, training and support through the critical adoption phase.

  7. Evolution

    A bespoke system does not end — it follows the company. Legislation changes, the operation grows, new needs appear.

Frequently asked questions

Frequently asked questions

How much does a bespoke system cost?

It depends entirely on scope, and any figure given before understanding the operation would be invention. What we can say is that we work in phases: the first delivery is sized to create value on its own, not to be a useless fragment until the project ends.

How long does it take?

A first working version usually takes two to six months, depending on complexity. Larger projects are split into phases with intermediate deliveries.

Who owns the source code?

That is settled in the contract, transparently, before starting. Our preference is that the client has certainty about the system’s continuity — including in the event of no longer working with us.

Do you work with companies in other states or countries?

Yes. We work remotely and serve clients in Brazil and Paraguay.

What if I already have a system and just want to improve it?

A common scenario. We assess what exists and recommend the path: evolve it, rewrite it in parts, or integrate. Rewriting everything is rarely the right answer.

Do you integrate with systems I already use?

Yes, and it is a speciality of the house — including closed or undocumented systems. See our Integrations page.

How does maintenance work after delivery?

A contract covering both corrective and evolutionary maintenance. A system in production needs continuous attention: legislation changes, volume grows, needs emerge.

Let us understand your case

The first conversation costs nothing and commits you to nothing. If we conclude that the best solution for you is not us, we say so — and point you where it makes sense to go.

Book a call