We specialise in tuning Java EE servers: Oracle WebLogic and IBM WebSphere.

Log in WebMail Jobs
B2B · CRM · ERP · Dashboards · Workflow

Custom web applications — when SaaS is no longer enough

Custom B2B portals, CRM/ERP systems tailored to your processes, BI dashboards, workflow automation, multi-tenant SaaS. Stack: Next.js / Java EE / PostgreSQL / Oracle. Integrations with Comarch, Subiekt, Sage, Salesforce. SLA 99.95%, contractual penalties, dedicated team.

01 / What we build

Six application types. Each built from scratch

You don't buy a "platform" with buttons that don't fit. Every application designed for your process, data, users. From MVP in 8 weeks to enterprise in 12 months.

B2B portals

Client-supplier portals, partner portals, distributor portals. Order placement, status tracking, invoices, documents, per-contractor negotiated prices, ERP integration.

login + roles · 2–6 months

Custom CRM

When Salesforce / HubSpot is too expensive or too rigid. Contacts, deals, pipeline, automation, email campaigns, reports, integration with company email, telephony, calendar.

3–9 months · multi-tenant ready

Custom ERP / ERP module

Warehouse, finance, HR, production, orders. Greenfield builds or a module extending Comarch / Subiekt. Approval workflows, reporting, JPK_VAT, e-Doreczenia.

6–18 months · legacy migration

BI dashboards

Real-time operational dashboards, KPIs for management, ad-hoc reporting. Integration with Tableau, Power BI, or own stack (D3, Recharts, Apache ECharts) under client branding.

2–5 months · live data

Workflow automation

Process automation: approvals, escalation paths, SLA tracking, integration with email, Slack, Teams. Instead of Excel + emails — a UI with timeline, audit trail, reports.

3–6 months · BPM

SaaS / multi-tenant

Subscription-based applications. Multi-tenant architecture, billing (Stripe), self-service onboarding, per-tenant admin, usage-based pricing, API for client integrations.

6–12 months · scalable
02 / Custom vs SaaS — honestly

SaaS for 80% of cases. Custom when it makes sense

Salesforce, HubSpot, Pipedrive, Microsoft Dynamics, monday — ready-made SaaS is good for most. Custom starts to pay off when SaaS limits you, license cost grows nonlinearly, or unique logic is a competitive advantage.

Ready-made SaaS (Salesforce, HubSpot, monday)

subscription · 50–500 PLN/user/month · ready in 1 week

When it makes sense

  • Standard business processes (sales, customer service)
  • Small team (up to 50 users) — licensing OK
  • No budget for custom build (50k+ PLN)
  • Need for quick start (1–2 weeks)
  • Compliance requirement that SaaS already has (SOC 2, HIPAA)

When SaaS stops working

  • Your workflow doesn't fit SaaS "best practices"
  • License cost for 500+ users = 250k+ PLN/year
  • Required integrations are expensive or impossible
  • Data must be on-prem (KNF, public sector)
  • SaaS doesn't allow your reporting

Decision rule: if 3-year license savings exceed the build cost (~150–500k PLN), custom makes sense. We'll help with TCO comparison before you decide — no fee.

03 / Tech stack

Four layers. Proven in production

Stack choice depends on the project. Simple CRM — Next.js + PostgreSQL. Enterprise application with integrations — Java EE + Oracle. High-traffic SaaS — serverless or Kubernetes.

Frontend

  • Next.js 14 (App Router, RSC)
  • React 18 + TypeScript
  • Vue 3 / Nuxt 3
  • Tailwind CSS / shadcn/ui
  • React Query / SWR
  • Zustand / Redux Toolkit
  • Recharts / D3 / ECharts

Backend

  • Java EE / Jakarta EE 10
  • Spring Boot 3 / WebFlux
  • Node.js + Hono / Fastify
  • Python + FastAPI / Django
  • NestJS (TypeScript)
  • REST + GraphQL APIs
  • WebSocket / Server-Sent Events

Databases

  • PostgreSQL 16 (preferred)
  • Oracle Database 19c / 23ai
  • MS SQL Server 2022
  • MySQL / MariaDB
  • MongoDB (schema-less)
  • Redis 7 (cache, queue)
  • Elasticsearch (search, logs)

Infrastructure

  • Oracle WebLogic / IBM WebSphere
  • Docker + Kubernetes
  • AWS / Azure / GCP / OCI
  • VMware vSphere (on-prem)
  • Nginx / HAProxy (LB)
  • Prometheus + Grafana
  • Sentry (error tracking)
04 / Process

Five stages. Sprints with demos

After every 2-week sprint you see working features on staging. No "big-bang launch". Iterations let you correct course before spending 200k PLN.

1

Discovery

Workshop 5–10 days: users, processes, integrations, compliance, SLA, target users. Output: Product Requirements Document.

1–2 wks
2

Architecture

System design, database ERD, API contracts (OpenAPI), security model, deployment topology, performance budget. Architectural Decision Records.

2–3 wks
3

MVP

3–5 core features. Auth, main workflow, basic reports. Production-ready (HTTPS, monitoring, backup), not a prototype.

8–12 wks
4

Iterations

2-week sprints with demos, feedback, prioritization. Client sees a working application from day 1, not "after a year".

3–9 months
5

Production + support

Launch to production, training, SLA monitoring, on-call rotation. Continuous development backlog. Documentation, runbooks, post-mortems.

ongoing
05 / Pricing models

Three billing formats. You choose based on risk

Time & Material for exploratory projects. Fixed Price when scope is clear. Dedicated team for long contracts. Mixed models (T&M to MVP, Fixed Price for development) are standard.

Time & Material

Daily rates, monthly reports, flexible scope.

daily rates · monthly report
  • Senior dev: 1,200–1,800 PLN/day
  • Mid dev: 800–1,100 PLN/day
  • Architect: 1,600–2,200 PLN/day
  • Monthly hour reports
  • Scope adjustable during the project
  • Ideal for R&D, MVP, exploration
  • Budget cap by agreement
Choose T&M

Dedicated team

Fixed team of 3–6 people for 12+ months, named.

12+ months · named team
  • 3–6 people (dev, QA, architect, PM)
  • Fixed monthly team rate
  • Fully dedicated resources
  • Daily standups + sprint planning
  • Direct client contact
  • Roadmap + backlog built together
  • SLA on new member ramp-up
Choose Team
06 / FAQ

Frequently asked questions

No answer? Write to us — an architect replies, not sales.

How long does a build from scratch take?
MVP (3–5 core features): 8–12 weeks from kick-off. Full production version with all features, integrations, reporting: 6–12 months. We iterate in 2-week sprints with demos on staging. First 2 sprints (4 wks) — architecture, auth, application skeleton. From sprint 3 you see working features on staging. Production-ready after 8–12 sprints (16–24 wks). Full maturity after 12–18 months (UX fixes, additional reports, extended integrations).
What SLA and uptime do you guarantee?
Depends on maintenance package. Standard 99.5% (43 hours downtime/year) — for most internal applications. Premium 99.95% (4 hours downtime/year) — for business P0. Enterprise 99.99% (52 minutes downtime/year) — for critical applications. SLA covers: (1) application availability (uptime monitoring, alerts), (2) P0/P1 response time (15 min — 4 h), (3) resolution time, (4) contractual penalties for breach. We host on Oracle WebLogic / IBM WebSphere in PROD or Kubernetes / AWS / OCI — depending on architecture. See Enterprise Support.
Who owns the code — client or you?
Client — as standard. After launch you receive full source code in your repo (GitHub, GitLab, Bitbucket), technical documentation (API, architecture, deployment, runbooks), access to all data. You can take the code at any time and continue with another team — we're not vendor lock-in. Standard contract: copyright and source code copies transferred to client on project completion or during (per release). NDA standard — we have a confidentiality obligation, client has full ownership.
What about 10k+ concurrent users?
Scalable architecture from day 1: (1) Stateless backend (horizontal scaling), (2) Database read replicas + write master, (3) Distributed cache (Redis cluster), (4) CDN for static assets (Cloudflare, AWS CloudFront), (5) Connection pooling (PgBouncer for PostgreSQL), (6) Background job queue (Bull, Sidekiq, RabbitMQ), (7) Auto-scaling Kubernetes / ECS Fargate. Performance budget: p95 latency <200 ms, throughput 1k+ req/s/node. Stress testing before launch (k6, Gatling). Real example: PZU application handling 50k users daily, p95 latency 87 ms.
Do we stop at MVP or continue developing?
As you prefer. After MVP you get a fully working production application. You can stop here (paying only for maintenance, ~590–2,990 PLN/month), or continue development in T&M, Fixed Price, or Dedicated Team model. Most clients continue — after MVP it's clear which features are critical, which integrations add value, where the application actually competes with SaaS. Development backlog agreed together each quarter.
How do I get started?
Most common path: (1) write to [email protected] with a description (application type, target group, scope, deadline, budget if set, whether you considered ready-made SaaS); (2) within 24 h an architect replies with 5–7 technical questions; (3) 30-min teleconference, optionally a pre-Discovery TCO comparison custom vs SaaS (free, 2 days); (4) offer — T&M / Fixed Price / Dedicated team with a specific price; (5) Discovery workshop 1–2 wks; (6) final SOW + MVP start. NDA + MSA + DPA as standard.
TALK TO AN ENGINEER

Consultation, audit, support

Tell us what you're facing. Your choices in the form build a brief — exactly what the engineer who replies will see.

How we work Audit, project, 24/7 care
Hours Mon–Fri 9–17 · outages 24/7 for clients on contract
  • we take over systems from other vendors
  • we support older WebLogic and WebSphere
  • urgent issue? quick diagnosis first

Tell us about it

01 I'm interested in…
02 Type of engagement…
03 Ballpark budget (PLN)…
04 Timeline…
05 Contact details…
PDF, DOC, TXT, JPG/PNG · max 5 MB
An engineer replies Not a salesperson, not a call centre — someone who understands your infrastructure.
A clear next step The first reply includes an initial scope and rough estimate.
Confidentiality & NDA on request Your environment details stay between us. We sign an NDA before the call if you need it.

Your data is safe. It will be used only to handle your request. More info in our privacy policy.