Skip to content
Thnkers

Custom SaaS development

From a one-line idea to a working system. In weeks.

We built a quoting system for a switchgear manufacturer in three weeks and an AI-review app for a school in two. Your problem is next.

What’s missing

You have an idea. You don’t have a dev team. Hiring one for a single product takes a year and kills the budget.

Bringing engineers in-house to build one product means recruiting, onboarding, management overhead and a salary commitment before a line of code exists. By the time the team is assembled, the window has often moved.

Contractors quote timelines in quarters and budgets that make the idea die on the spot.

You describe the problem. They come back with a discovery phase, a scoping phase, a design phase and a development phase. The estimate is six months minimum, on a number that assumes you’ve already de-risked everything. You haven’t. That’s why you came to them.

Investing in development before validating demand is frightening. It should be.

Building the wrong thing is expensive in ways that take a long time to become visible. A year of development, a launch, then silence. The fear is reasonable. The solution isn’t to move slower — it’s to validate first.

You already have a zoo: bots, spreadsheets, no-code glue. It works, but it’s fragile and embarrassing to show anyone.

Three Telegram bots holding hands. An Airtable only one person understands. A Zapier chain that breaks whenever someone changes a field. The operational logic is real and often sophisticated. The infrastructure holding it together is one update away from an incident.

The promise

We turn “is it possible to make it so that…” into a working system in production — in weeks, not quarters. Not a demo, not a prototype: roles, authentication, deployment, and real documents going out to your clients. We find the actual problem first, then build the solution to that — not to the original request.

What we do

  • Product discovery

    We start with a conversation, not a brief. What’s breaking, what’s slow, what the person who knows the operation best would fix first. The answer is usually different from the initial request. The switchgear client asked for AI. What they needed first was a system. We found that in the first conversation, not six weeks into a requirements document.

  • MVP in two to four weeks

    A minimum viable product in production, on real data, used by real people. Not a staged demo, not a clickable prototype. The kind that produces an actual document your client receives, an actual grade a student sees, an actual payment that clears. The only kind worth building.

  • Full cycle, one team

    Frontend, backend, database, integrations, authentication, deployment. The person who found the real problem writes the specification and the same team builds the system. Nothing is handed off. Nothing gets lost between a consultant who scoped it and a team that builds it.

  • Automation where it removes real manual work

    Reading engineering drawings and extracting component lists. Checking homework against structured criteria. Parsing documents and matching items to a catalogue. Things that take a person’s time and don’t require human judgment — automated with the right tool, and a human confirmation step where it matters.

  • Migration from fragile solutions

    If the operational logic lives in bots, spreadsheets or no-code, we move it into a single product. The principle: change the interface and the infrastructure, keep the business rules exactly as they are. The Alfa Homework migration worked because we ported the domain logic from three bots into one API without changing what it does.

  • Iteration and support

    The product grows from what real usage shows, not from what the initial spec assumed. We offer a retainer, but the product is designed so you can take it to any team. The code, the repository, the documentation and all credentials are yours from day one.

Compliance by design

For AI work in the EU we build to the AI Act from the start — risk classification, decision logging, data documentation and human oversight, with the technical file included in delivery.

How we work

Conversation first, not a brief. We diagnose the actual bottleneck before agreeing a scope. Then we fix the MVP scope, build with weekly demos so nothing is a surprise, and ship to production. After that, we iterate based on what real users do with it. The code and the data belong to you.

Where we have done it

Questions

  • How much does an MVP cost?

    The range depends on scope. After the first conversation we give a fixed price for discovery and an MVP estimate. No “we’ll assess as we go.”

  • Who owns the code?

    You do. Repository, documentation and all access credentials transfer to you completely.

  • What stack do you use?

    React and TypeScript on the frontend, Node and PostgreSQL on the backend, React Native and Expo for mobile. Reliable, well-documented, and easy to hire for if you bring development in-house later.

  • What happens after you ship?

    We offer a support and development retainer, but the product is designed so you can take it to any team. We don’t make ourselves hard to leave.

  • We don’t know how to write a technical specification.

    You don’t need to. We start with a conversation. Writing the specification is our job.

Start a project

Describe your problem in one sentence — the way the switchgear client did. We’ll tell you what we’d build and what it would cost.

Start a project