Custom Software Development
When the SaaS runs out, build the thing that fits the business.
There is a point where configuring a platform costs more than building the thing. It usually arrives when the workaround needs its own workaround, or when the per-seat licence for a tool nobody likes crosses the cost of owning the alternative.
We build production software: multi-module ERP systems, customer and partner portals, internal tools, installable progressive web apps. Typed end to end, permissions enforced in the database rather than only in the interface, and deployed somewhere you control.
The bias is towards boring, durable technology (Postgres, Next.js, row-level security) because the interesting part should be your domain, not the stack.
What you end up with
Systems that fit the process
Modules shaped around how the business actually runs (its approval chain, its document types, its reporting) instead of the shape a generic tool imposes.
Permissions that hold
Roles as a grid of modules against actions, enforced at the route, in the write layer, and again in the database, so no screen can forget.
Nothing is really deleted
Soft delete, a restorable trash, and an append-only activity log with field-level diffs. Only an administrator can destroy anything permanently, and the audit entry survives that too.
Installable and offline-capable
Progressive web apps with their own home-screen icon, full-screen launch and a cached shell that still opens when the signal drops.
What actually lands
Shaped to the engagement, but this is the shelf it comes off.
- 01Next.js applications with typed data access and server-side rendering
- 02Postgres schema design with row-level security and migrations
- 03Authentication, role-based access control and per-user permission overrides
- 04Audit logging, soft delete and approval workflows
- 05Reporting, CSV and PDF export, and dashboards
- 06PWA packaging, generated icon sets and offline shells
Where this has shipped
Custom development, specifically
Buy, wherever a product genuinely fits. It is nearly always cheaper. Building earns its place when the process is your actual differentiator, when licence cost per seat has overtaken the cost of ownership, or when the integration work to make three tools behave as one has become larger than the thing itself.
You do, in your own repository, from the first commit. Deployment runs on your accounts (your Vercel project, your Supabase project, your domain), so there is no vendor lock to us. If you take the work in-house, nothing needs to be untangled first.
A handover that covers the schema, the deploy path and the runbook, plus whatever ongoing support arrangement suits. Plenty of clients take it fully in-house; some keep a retainer for changes. Both are fine, and the code is written on the assumption that someone else will read it.
Page copy drafted with AI assistance. Skofi is accountable for it. Editorial policy
Where this is applied
Tell us about the process that keeps costing you.
A 15-minute call is usually enough to say whether this is an automation, an integration, or something that should be built, and whether we are the right people for it.
What happens on the call
- 15 minutes, no deck
- We map the process on the call
- A written scope, or an honest no