OUR OFFER - CREATING WEB APPLICATIONS

Web application for a company that has outgrown its sheets

Custom CRM, customer dashboard, B2B portal, internal dashboard, SaaS multi-tenant - when off-the-shelf tools (Excel, Pipedrive, Asana, Zapier) stop scaling with your business. Application designed for your workflow, not for generic SaaS assumptions - expandable iteratively over 3-5 years, no need to refactor after a year.

Most companies start with off-the-shelf tools: Excel, basic CRM, Zapier for integration. It works up to the scale of a few dozen people and a few processes. Above that scale, the reverse economy begins - more hours of manual data copying, more operational errors, more tools that no one uses to their full potential.

A custom web app solves what off-the-shelf tools don't address: workflow specific to the customer's industry, integration between 5+ systems in the company, customer dashboard with self-service 24/7, management dashboards aggregating data from the entire ecosystem. Most of our web application clients are companies of 30-150 people - Excel stopped being enough 6 months ago, but off-the-shelf SaaS doesn't fit the process or costs more than a custom build with a 5-year horizon.

20+ years of IT - 50+ implementations - from PLN 30,000 - Poland / EU / Switzerland

OUR OFFER - CREATING WEB APPLICATIONS

Web application for a company that has outgrown its sheets

Custom CRM, customer dashboard, B2B portal, internal dashboard, SaaS multi-tenant - when off-the-shelf tools (Excel, Pipedrive, Asana, Zapier) stop scaling with your business. Application designed for your workflow, not for generic SaaS assumptions - expandable iteratively over 3-5 years, no need to refactor after a year.

20+ years of IT - 50+ implementations - from PLN 30,000 - Poland / EU / Switzerland

WHEN IT'S WORTH IT - 4 SIGNALS

When a web application pays for itself faster than an off-the-shelf SaaS

Four signals at which custom application is cheaper long-term than trying to squeeze the process into off-the-shelf tools. If you recognize at least two - it's time for an architectural conversation.

01WORKFLOW DOESN'T FIT

Implemented SaaS forces process change instead of support

You bought an off-the-shelf CRM, project management or industry SaaS - the team complains. The company's workflow doesn't match the tool's assumptions, so either Excel comes back through the back door or the process has been twisted to fit the tool.

Specific symptoms:

Custom workflow requires 5+ workarounds in the finished tool
30%+ of functions not used because they don't fit the industry/process
Employees say "simpler to do it in Excel"
Subscriptions to automation tools are on the rise because SaaS-ready lacks key integrations

Custom application: workflow 100% under the company's process, one database, one login, one logic.

02DATA SILOS

Data between systems flows through people, not APIs

Order from marketplace copied manually to accounting. Lead from web form transcribed into CRM. Invoice re-entered into management report. Each operation is 5-15 minutes

risk of error.

Specific symptoms:

3+ systems in the company, data flows through Excel/email/PDF
Backoffice worker spends >20% of time manually copying data
Management reports are late because data needs to be "downloaded"
Each new customer = setup in 4-5 systems separately

Custom application with API integration: real-time synchronization, single source of truth, management reports automatic.

0324/7 SELF-SERVICE

B2B customer expects self-service, we serve by email

The B2B customer wants 24/7 access: order status, invoice history, downloadable documents, claim requests. The competition already has this. You handle email - speed goes down, costs go up, the customer sees you as "less professional."

Specific symptoms:

The customer asks "do you have a customer panel?"
30%+ of service team's time is repetitive status questions
Returns/complaints generate 5+ emails back and forth
Competitors in the industry have a customer portal, you do not

Custom customer panel: reduce repetitive service by 50-70%, signal market professionalism.

04DECISIONS ON DATA

Management decisions based on late Excel reports

Management gets monthly reports - it takes 3-5 analyst working days to prepare. Data is late, decisions a week too late. Real-time dashboard with aggregation from CRM, accounting and warehouse solves the problem.

Specific symptoms:

Management reports prepared manually, 1-3 days of work per report
Management asks "how are we doing?" - the answer requires 4 emails to various departments
Data from 3+ systems combined in Excel from scratch each time
Analyst employee overloaded with work, most of which is repetitive

Custom dashboard: real-time data with aggregation, self-service drill-down, alerts for anomalies.

↳ If you don't recognize any of the 4 signals - off-the-shelf CRM / project management / SaaS will probably suffice and are cheaper than a custom application. We're stating it plainly instead of pushing an unnecessary investment.

5 SCENARIOS - CHOOSE THE TYPE OF APPLICATION

What specifically are we building

Each scenario is a separate sub-page with details: when worth, typical usage, stack, process, price fork. SaaS as a flagship - broader scope, generic case. The other 4 are specializations for specific business needs.

01SAAS

SaaS multi-tenant - from MVP to product

Multi-tenant - Subscriptions - Auth - White-label - Scaling-ready

Widest scope in web applications - generic case for companies that are building their own SaaS product or an internal application that works like SaaS for multiple departments. Multi-tenant architecture from day one, subscriptions (Stripe), auth with multi-factor, optional white-label for resellers.

Stack: Next.js + Payload + PostgreSQL + Stripe. Modular architecture under 3-5 years of iterative development.

Typical project: 8-16 weeks. Price: from PLN 50,000.

02DASHBOARD

Management dashboard with real-time data

BI - Data aggregation - Self-service drill-down - Alerts.

For management and operations departments needing real-time insight into company data. Aggregation from CRM, accounting, warehouse, production. Self-service drill-down instead of manual Excel reports. Alerts for operational anomalies. Eliminates 3-5 days of analyst work per month.

Typical project: 6-12 weeks. Price: from PLN 30,000.

03DASHBOARD

Management dashboard with real-time data

BI - Data aggregation - Self-service drill-down - Alerts.

Typical project: 6-12 weeks. Price: from PLN 30,000.

04CUSTOM CRM

Custom CRM for industries not served by off-the-shelf SaaS

Workflow - Pipeline - Integrations - Industry-specific

When Pipedrive, HubSpot and others don't fit your specific industry workflow (medical, finance, B2B niche, manufacturing, logistics). Custom pipeline, custom stages, custom bid calculations. One base, one login, integrations with existing company systems.

Typical project: 8-14 weeks. Price: from PLN 40,000.

Dedicated sub-site in preparation - full description and calculator available soon.

05B2B PORTAL

B2B partner portal for companies of 100-300 people

Partners - Distributors - Agents - White-label - Multi-region.

For companies growing in scale of partner distribution - portal for agents, distributors, regional partners. Each partner can see his data (orders, margins, marketing materials), full isolation per tenant. White-label for co-branded operations.

Typical project: 12-20 weeks. Price: from PLN 80,000.

Dedicated sub-site in preparation - full description and calculator available soon.

↳ The list is not closed. If your case does not fit any scenario (custom marketplace, internal industry platform, employee tool) - describe in the form, we come back with specifics in 24-48h.

WHAT YOU GET - 4 APP-SPECIFIC CONCRETES

What do you get in the price of a DV web application

Four things in the standard of any DV web application - no matter whether SaaS, customer portal, dashboard or custom CRM. Not upsells, not premium packages - the baseline for every project.

UX DESIGNED FOR BUSINESS

UX tailored to real-world workflow, not to design trends

Most web applications have UX"copied from other SaaS". - sidebar nav, dashboard tiles, generic forms. It looks modern, but doesn't match the client company's workflow. Employees click more, not less.

DV works differently:

User research with 3-5 prospective users before design begins (Stage 02 of the process) - what they realistically do, how long, how often
Wireframes tested with users before code (clickable Figma prototype)
Design tokens as CSS variables - visual consistency, but workflow-first
A/B testing of the most important post-launch flow - iterating on data, not guessing

The consequence: employees adopt the application in 1-2 weeks, rather than"we have been learning this for 3 months"..

MONITORING AND ALERTS FOR THE OWNER

You know sooner than the user when something doesn't work

A web application is different from a website - users are with it every day, sessions last hours, mistakes cost real money (lost order, bad calculation, lost data). Monitoring is not"nice to have", is the baseline.

DV standard in every web application:

Sentry - error tracking with full stack trace + session replay (we can see what the user was doing before the error)
Uptime monitoring every 5 minutes - SMS/email alerts when application unavailable >2 minutes
PostHog - product analytics + heatmaps + session recordings
Application Performance Monitoring (response time, DB queries, bottlenecks)
Centralized logging with retention 30 days + alerts for anomalies

Consequence: production error report We to you. Fix typically in 2-4h from incident. Your team does not waste time on"someone says it doesn't work, let's check"..

ARCHITECTURE UNDER 3-5 YEARS OF DEVELOPMENT

An application to which functions are added rather than rewritten

Web applications live for 3-5 years. Architectures designed for this perspective, not for"by launch in 8 weeks".:

Monorepo structure (apps/web, apps/admin, packages/database, packages/ui) - add admin panel or mobile app without rewriting
TypeScript strict mode + Zod validation - contracts between modules checked compile-time (errors caught before deployment)
Database migrations in versioned code - schema evolution without"ALTER TABLE in production".
Multi-tenancy ready (even single-tenant applications) - adding a second client is a configuration, not a refactor
Feature flags + A/B testing infrastructure from day one

Consequence: expansion in 2 years is adding features in 2-4 sprints, not a 6-month refactor of the entire system.

AUTH AND SECURITY AS STANDARD

Security invisible to the user, but present

The web application iscustomer data of your company - A breach means a real legal (RODO), reputational and operational problem. Security baseline as standard:

Authentication: NextAuth/Auth.js or Clerk (multi-factor 2FA, passwordless, SSO optional)
Authorization: role-based access control (RBAC) from the code and database level (defense in depth)
Encrypted at rest (PostgreSQL TDE) + encrypted in transit (TLS 1.3)
OWASP Top 10 covered: SQL injection, XSS, CSRF, broken auth, security misconfiguration
Rate limiting on API endpoints (protection against brute-force and DDoS)
Audit log at the base: who, when, what changed (RODO compliance)

The consequence: the application passes the enterprise client's security audit without panic refactoring. The client's compliance team gets documentation (security checklist) in the handoff.

↳ These 4 elements are included in the price of a DV web app from the first project - not a "premium package". Software house offering them as an upsell is a red flag - in a web application (vs website) they are a baseline, not an add-on.

STACK - 4 LAYERS OF ARCHITECTURE

Technology stack for web applications

Web application stack selected according to business requirements -. there is no "default stack" in DV. Four layers below (frontend, backend, auth, infrastructure) with the most common choices and criteria for decisions per project.

01FRONTEND

Next.js + React Server Components + Tailwind

Default frontend for web applications can be: Next.js with App Router and React Server Components. Tailwind CSS + shadcn/ui as component library. React Hook Form + Zod for form validation.

Why this stack:

Next.js App Router - server components reduce JavaScript bundle 30-60% vs pure client React, better LCP/TTI
Tailwind + shadcn/ui - design tokens as CSS variables, component library copy-paste (not node_modules dependency)
Zod validation - single source of truth for types frontend + backend + database

When to withdraw:

Vue / Nuxt - when the customer already has the Vue ecosystem in-house
Astro - for content-heavy applications with minimal interactivity (documentation, information portal)
Angular - enterprise customers with existing Angular database

In 90% of DV web application projects: next.js. The remaining 10% depends on the client.

02BACKEND AND DATABASE

Node.js + TypeScript + PostgreSQL + Prisma

Backend for web applications can be implemented based on: Next.js API Routes (serverless functions) or dedicated Hono/Fastify for high-throughput APIs. Database: PostgreSQL with Prisma ORM. Background jobs: BullMQ

Redis (when needed) or Inngest (serverless background tasks).

Why PostgreSQL:

ACID transactions - zero risk of data loss
JSONB columns - flexibility NoSQL, but in a relational database
Row-level security - multi-tenancy ready without application-layer hacks
Production proven - since 1996, used by Stripe, Reddit, Instagram

Why Prisma:

Type-safe queries - error in queries caught compile-time
Auto-generated migrations with schema diff
Studio (GUI for base) in development, prod-safe in production

Alternatives considered: Drizzle (lighter ORM, edge-deployment), TypeORM (legacy projects), raw SQL (when Prisma constraint).

03AUTH AND MULTI-TENANCY

Auth selected for the application's business model

Auth does not have one"default'a" In web applications - depends on the business model:

Single-tenant (one company, many employees):

NextAuth/Auth.js + email magic links or OAuth (Google, Microsoft) - simplest, zero vendor lock-in
Role-based access control from the base (RBAC) - admin/user/viewer

Multi-tenant SaaS (multiple clients, each with their own workspace):

Clerk or WorkOS - turnkey solutions with multi-org support, SSO, SCIM provisioning
Custom solution on NextAuth (when budget tight or specific requirements)

B2B portal with partners:

SSO (Single Sign-On) from the client system - SAML 2.0 or OAuth
Provisioning automatic from HR of the customer's system

Auth security baseline (each application):

Multi-factor 2FA optional for user
Rate limiting on login endpoints (brute-force protection)
Audit log of each login (RODO compliance)
Sessions with secure HttpOnly cookies

04 INFRASTRUCTURE AND DEPLOYMENT

Infrastructure selected for scale and requirements

The most common DV web application deployment scenarios:

VPS - 80% of projects of Polish clients:

VPS - cost 80-200 PLN/month.
Docker containers, Traefik load balancer, automated SSL
Database PostgreSQL on the same VPS or separate (depending on scale)
Backup daily + retention 30 days

Cloud (AWS / GCP / Azure) - 15% of projects:

When the customer already has a cloud ecosystem (Azure AD, AWS S3, GCP BigQuery)
Serverless architecture (Lambda/Cloud Functions) for event-driven workflows
Managed database (RDS/CloudSQL) - higher cost, but managed by cloud provider

Vercel + Neon/Supabase - 5% of projects:

Global edge network - when an app has users on 3+ continents
Zero DevOps overhead - automated preview deployments per PR
Trade-off: cost increases non-linearly with scale (>200k MAU)

CI/CD any application:

GitHub Actions - automated tests + build + deploy for staging after each merge
Production release manual after code review + acceptance testing

↳ Stack opinionated. Changing the default (e.g., Vue instead of React, Cloud instead of VPS) requires justification in Stage 01 of the process - we don't change the stack "because the customer prefers it" without analyzing the long-term consequences.

PROCESS - 5 STAGES - 8-16 WEEKS5 steps

From consultation to production - in stages

Each DV web application project goes through 5 stages. The duration depends on the scale (8 weeks for a customer portal, 16+ for SaaS multi-tenant), but the structure is constant. Decisions architectural decisions are made at stages 1-2 - not after the fact. *"We forgot about auth multi-tenant "*.

1-2 weeks - free of charge
ARCHITECTURAL CONSULTATION
First call (60-90 min) + 1-2 follow-up sessions. We map: what business processes the application is supposed to support, what systems exist in the company (CRM, ERP, accounting), what the future functionality roadmap looks like (3-5 years). Build/buy decision - we verify if custom actually makes sense vs ready-made SaaS.
Output: ADR (Architectural Decision Record) v0 + stack decision (frontend / backend / database / auth / infrastructure) with 2-3 alternatives per layer. Each decision has a business case.
2-3 weeks - paid deliverable
DISCOVERY AND SPECIFICATION
Discovery deep-dive specific for web application: user research (3-5 interviews with prospective users), mapping data model (how application organizes data), API integration specification (which external systems, which endpoints, sync frequency, error handling), security model (auth, RBAC, multi-tenancy isolation if SaaS).
Output: 4 heritage tabular documents:
ERD (Entity Relationship Diagram) - database schema.
API contract (OpenAPI 3.0 spec).
Integration maps - which external systems, which operations
User stories + acceptance criteria - testable requirements
For SaaS multi-tenant additionally: tenant isolation strategy (row-level security vs schema-per-tenant vs database-per-tenant).
2-4 weeks
UX/UI + PROTOTYPE TESTED
Wireframes low-fidelity for all key user flows (login, major CRUD operations, reports). Then design tokens and high-fidelity mockups in Figma. Clickable prototype tested with 3-5 usersprior to encoding - We identify 60-70% of UX issues before they become tech debt.
For applications with existing design system of the client (after branding DV or other supplier) - we adapt to existing guidelines instead of creating from scratch.
Clickable Figma prototype + design tokens as CSS variables ready for import in Tailwind config. Zero hand-translation design → code. User testing before coding = reduced re-work by 40-60% in Stage 04.
4-12 weeks - 2-week sprints
DEVELOPMENT + STAGING FROM DAY 1
Longest stage. We work in 2-week sprints with a review after each sprint.Staging environment from the first sprint - Customer tests N+1 feature while working on N+2, feedback loop 1 week instead of 3-month big-bang launch.
MVP scope: core user journey + critical integrations. Secondary features (advanced reports, edge case handling, admin panel deluxe) - after MVP launch based on real user feedback.
Continuous deployment to staging. Code quality in CI/CD: TypeScript strict, ESLint, Vitest unit tests, Playwright E2E for critical paths (login, payment, form submit, API contract). Merge blocked if not passing standards.
Each PR = preview deployment with unique URL - client tests changes isolated from main staging.
Launch + 30 days of stabilization
LAUNCH + MONITORING + STABILIZATION
Production deployment +monitoring stack configured before launch (Sentry, Uptime, PostHog, APM, centralized logging). Handoff session 30-45 min: architecture, deployment process, how to debug production, how to push changes. 30 days of stabilization included (minor fixes, optimizations, bug fixes discovered in production).
After stabilization 3 options:
End of cooperation - zero lock-in, code yours
Technical care subscription
Expanding more features
Handoff documentation heritage tabular:
README + setup instructions (local dev in 30 min)
ADR (Architectural Decision Records) - why this stack
Deployment runbook (how to deploy, how to rollback, how to debug production)
Database schema docs + migrations history
Integration runbooks (how to set up Stripe webhooks, how to rotate API keys, how to debug failed sync)
Playbook monitoring - what alerts mean what, when wakeup-on-call (if critical application)

WEB APPLICATION PROJECT SCHEDULE

8-16 weeks

TOTAL

8-12 weeks

CUSTOMER PORTAL / DASHBOARD

12-20+ weeks

SAAS MULTI-TENANT

↳ Specific schedule after Discovery (Stage 02). SaaS enterprise projects (multi-region, compliance audit, dedicated support) - 20+ weeks. We consult individually.

PRICE LIST - 3 TIERS OF WEB APPLICATION

How much does a web application cost

Each project is priced individually after Discovery. Below 3 tiers of scale typical for DV web applications + calculator for estimating your own project in 4 minutes.

01LITTLE

30 - 60 thousand zlotys.

Time: 6-8 weeks

Single module, 1-2 user roles

1-2 user roles (e.g. admin + user)
5-15 core screens
1-2 API integrations (e.g. Stripe + email)
Single-tenant architecture
Auth standard (NextAuth/Clerk)
Monitoring stack (Sentry + Uptime + PostHog).

02AVERAGE

60 - 120 thousand zlotys.

Time: 8-12 weeks

3-5 roles, integrations, simple multi-tenant

3-5 user roles with RBAC
15-40 core screens
3-5 API integrations (CRM + accounting + payments + email + analytics)
Multi-tenant ready (single-tenant deploy, but architecture for scaling)
Auth with multi-factor + audit log
Background jobs (queue processing, scheduled tasks, email sending)
Full monitoring + alerts

03COMPLETED

120 - 300+ thousand zlotys.

Time: 12-20+ weeks

Multi-tenant SaaS, compliance, advanced features

Multi-tenant SaaS architecture (row-level security or schema-per-tenant)
5+ user roles with complex RBAC + custom permissions
40+ core screens
5-15 API integrations + bidirectional webhooks
Advanced features: white-label, custom fields per tenant, workflow engine, reporting builder
Compliance ready (RODO Art. 9, FSC, audit trail)
Real-time collaboration (WebSocket/SSE).
Subscriptions (Stripe billing) + usage-based pricing

What most strongly determines the price of a web application

Honest Scope:

Hosting and infrastructure in production (VPS 80-200 PLN/month, cloud per usage, dedicated from 500 PLN/month).
Monitoring tools - Sentry, PostHog, Uptime monitoring (Free for small scale, further from ~$25/month per tool)
External services paid per use (Stripe 1.4-2.9% per transaction, Twilio per SMS, OpenAI per token, email delivery from PLN 50/month).
Domains + SSL (~£100/year)
Campaign marketing (App Store Optimization, ads, content)
Content (copywriting, product graphics) - we can identify proven providers

Long-term costs: usually 8-15% per year of the project value for hosting + technical care + minor extensions.

What our forks don't cover

Heritage scope - what's outside our area of work:

Lean MVP for 10-20 thousand zlotys. - At this budget, a throwaway code is usually created. Better options: no-code (Bubble, FlutterFlow) as first validation, senior freelancer with portfolio, startup gas pedal with built-in tech support.
Custom enterprise with budgets of £1mil. - multi-region failover, custom hardware, dedicated support team 24/7 - fit larger consulting firms (100+ person software house).
Migration of legacy enterprise systems (mainframe, JEE legacy, custom ERPs 15+ years old) - DV does not work with legacy stack.
Single fixes in another agency's existing application - We work on a project basis (Discovery → Deployment → Long-term care), not hourly for minutiae.

If your project falls into the above categories - we say so directly instead of taking a project that we will not do optimally.

What is NOT included in the project price

Honest Scope:

Hosting and infrastructure in production (VPS 80-200 PLN/month, cloud per usage, dedicated from 500 PLN/month).
Monitoring tools - Sentry, PostHog, Uptime monitoring (Free for small scale, further from ~$25/month per tool)
External services paid per use (Stripe 1.4-2.9% per transaction, Twilio per SMS, OpenAI per token, email delivery from PLN 50/month).
Domains + SSL (~£100/year)
Campaign marketing (App Store Optimization, ads, content)
Content (copywriting, product graphics) - we can identify proven providers

Long-term costs: usually 8-15% per year of the project value for hosting + technical care + minor extensions.

Count the cost of a web application Describe your project

↳ Forks based on 50+ DV web application deployments over the last 3 years. Final price ±15% after Discovery (Stage 02). We consult individually on SaaS enterprise projects (multi-region, industry compliance, custom hardware).

WHAT MAKES US DIFFERENT - 3 APP-SPECIFIC THINGS

3 things we do differently in web applications

Regardless of the tier (small / medium / complex) three things below are included in the price and standard of every DV web application. A software house offering them as a *"premium add-on "* in a web application web application (vs. website) is a red flag.

MULTI-TENANCY READY

Multi-tenant architecture even for single-tenant projects

Most software houses build single-tenant applications"because the customer has one company".. Consequence: when a customer in 18 months wants to add a second customer, a second brand, a second region - a 6-month refactor of the entire database and auth system.

DV works differently. Any web application designedmulti-tenant ready from the first sprint, even if the deploy is single-tenant:

Database schema withtenant_id in each table (even for single-tenant - tenant_id = '1')
Row-level security (RLS) in PostgreSQL - isolating data at the database level, not the application level
Tenant routing (subdomain or path-based) - ready for activation in 1-2 sprints
Per-tenant config (logo, primary color, feature flags, pricing plans) - schema ready, UI panel added on-demand
Middleware auth with tenant context - request always knows its tenant

The consequence: adding a second customer in 2 years isconfiguration, not refactor. White-label for resellers = 1-2 sprints of work, not 6 months. Mostly used for single-tenant SaaS → multi-tenant conversion after first enterprise customers.

PRODUCTION-READY

Staging from sprint 1, pre-launch monitoring, zero "big bang"

Typical software house: development in isolation,"we are showing the MVP in 8 weeks".. The customer sees the app only after launch, catching issues in production, panic mode for the first 2-4 weeks after go-live.

DV works differently:

Staging environment from sprint 1 - link current always, client is testing feature N+1 while working on N+2
Continuous deployment - any merge to main = auto-deploy to staging in 5-10 mins
Preview deployments per PR - each pull request = unique URL with isolated deployment
Code quality in CI/CD - TypeScript strict, ESLint, Vitest unit tests, Playwright E2E for critical paths. Mergeblocked if it doesn't pass standards
Monitoring stack configured before launch - Sentry, Uptime, PostHog, APM, centralized logging. Not"we will add after the launch"., only"works before the first user"

Consequence: launch isnon-event. Customer tested app 8 weeks before go-live, monitoring catches errors before user notices, fix typically 2-4h from incident. Zero panic mode in the first weeks of production.

CODE OWNERSHIP

Code in your repo from the first sprint, zero proprietary DV libraries

Vendor lock-in in web applications is a bigger problem than in websites. An application is the foundation of a company's operations - the cost of changing developers can be 50-100% of the cost of re-implementation. Software houses, knowingly or unknowingly, create dependencies that are difficult to break.

DV works differently from the first sprint:

Full source code in your GitHub/GitLab - From first commit, not after launch, not after last invoice
Your license, your intellectual property - zero clauses"DV retains the rights to the code". contract
Open-source / standard frameworks only - Next.js, Payload, PostgreSQL, Stripe - zero proprietary DV libraries that are not on npm
Zero hidden integrations - each third-party integration has a trail audit in the code + documentation on how to configure it under the client's own account
Handoff documentation included - README + ADR + deployment runbook + database schema docs + integration runbooks (6 heritage tabular documents)

Consistency: a client with an active in-house developer will find himself in code in 1-2 weeks. A client changing developer/agency after 2 yearsdoes not lose months at"decryption". Real case from our practice: 2 clients switched to in-house team after 18-24 months - code moved without problems.

↳ These 3 elements are included in the price of a DV web application from the first project. Software house offering them as "enterprise tier" or "premium add-on" - a signal that they are not a priority for the team and the client is paying separately for things that should be standard.

FOR WHOM DV - AND FOR WHOM NOT

Who we build web applications for - and who we don't

DV works well with 3 types of web application clients. For the 3 types of projects we recommend elsewhere - a ready-made SaaS, a freelancer or a larger software house. Honest scope filter before we start discussion.

Company of 30-150 people, custom system as a must-have

Your sweet spot at DV. Excel/ready CRM ceased to suffice 6+ months ago, company traffic requires dedicated application (custom CRM, customer panel, B2B portal, internal dashboard, SaaS for internal departments). Budget 50-150k for first project, decision cycle 2-6 months.

Typical industries: professional services (legal, medical, financial), B2B SaaS for niche markets, e-commerce with custom workflow, manufacturers with exports, service companies with location networks.

What we give you: dedicated application under your workflow, multi-tenant ready architecture from day 1, monitoring and long-term technical care (450-1500 PLN/month).

Ready-made SaaS covers 70%+ of your needs

Custom web application islong-term investment (PLN 50-300k + 5-15% per year maintenance). If a ready-made SaaS covers most of your needs - buy, don't build.

The most common situations when we say "buy SaaS ready":

CRM for 10-30 people - Pipedrive, HubSpot Free, Bitrix24 starter
Project management - turnkey project management SaaS, industry SaaS (e.g., for architects, lawyers).
Mailing automation - Brevo, Mailchimp, ActiveCampaign
Booking/reservation - Calendly, Cal.com, industry SaaS
Helpdesk/ticketing - turnkey solutions from ~£200/month.

In a 20-minute architectural consultation, DV will walk you through a build/buy decision. 15-20% of the conversations end with the recommendation "don't build, buy ready-made SaaS X".

Startup post-MVP, refactor or rebuild under scale

You have a validated business case, first revenues, but an MVP built on no-code (Bubble/FlutterFlow) or in throwaway code is starting to crack. You need a production-ready system under scale (100+ active users, multi-tenant, billing).

Budget 80-200k for refactor/rebuild + iterative expansion plan. Decision cycle 1-3 months. Series A in 6-12 months = application must be due-diligence ready (code quality, documentation, security baseline).

What you get with DV: production-ready stack from day one, multi-tenant architecture, monitoring + alerts, due-diligence documentation in handoff.

Lean MVP for 10-20 thousand zloty / proof-of-concept in a week.

If your budget is PLN 10-20 thousand and you need an applicationto validate the business hypothesis (not to make production work for 100+ clients) - DV is overkill. Our minimum project value is 30K (Tier 1 small app), because below that budget we won't build an app that makes long-term business sense.

Recommended alternatives:

No-code platform - Bubble (strongest for web apps), FlutterFlow (mobile + web), Glide (simple internal tools)
Freelancer senior With a portfolio of similar projects (Useme, Bulldogjob, Polish tech communities)
Startup gas pedal with built-in tech support (PARP Startup Platforms, Startup Booster Poland - funding)

Come back to us when the MVP validates the business hypothesis and you need a production-ready system under scale (100+ users, multi-tenant, SLA, compliance).

Company of 100-300 people, outsourcing a module of a larger system

You have an internal tech team, but overloaded - looking for a partner for a dedicated application module (partner portal, booking system, B2B marketplace, custom integration layer for existing ERP). Budget 100-300k, decision cycle 3-9 months with procurement process (RFP, NDA, DPA).

What you get with DV: an external module designed for your existing architecture, handover-ready technical documentation for your in-house team, code in your repo, a contract with enterprise compliance clauses.

Enterprise 500+ employees, budget 1+ mil, multi-region

For enterprise projects with requirements: multi-region failover (deployment in 3+ regions), custom hardware integration (IoT, embedded, dedicated GPU), dedicated support team 24/7 with SLA 99.99%, compliance audit (SOC2, ISO 27001, HIPAA), advanced ML/AI infrastructure with proprietary models - DV is too small a team.

Fit larger consulting firms (100+ people software house Polish/international) or direct hire enterprise dev team.

Our sweet spot: web applications up to 300K budget, teams up to 300 people, scope 1-2 core systems. Above this scale, we work with youonly as a supplier of a specific module - not the main contractor for the entire platform.

↳ If you fit into one of the 3 profiles on the left - the next step is an architectural consultation (60-90 min, free of charge). If you fit into 1 of the 3 situations on the right - thank you for your time on this page. We recommend ourselves on the recommendation of companies we have worked with or reviewed in the market.

FAQ - 7 QUESTIONS ABOUT WEB APPLICATIONS

Questions we get during the consultation phase

Questions specific to web applications - multi-tenancy, scaling, security, compliance, refactor with legacy. Questions general about working with DV - in /our-offerings FAQ block. Questions about software in general - in /creating-software FAQ block.

01MULTI-TENANCY
When should an application be multi-tenant, and when is single-tenant enough?

Three decision criteria:

Will the application support >1 customer/team/brand in the future 3-5 years?

Yes (even hypothetically) → multi-tenant ready from day 1
Not sure → multi-tenant ready architecture, but single-tenant deploy (DV default)
Definitely not → single-tenant (rare in practice)

Is the app ever going to be white-label for resellers?

Yes → multi-tenant required, custom branding per tenant
No → single-tenant ok

Compliance requirement to isolate data between customers?

Yes (medicine RODO Art. 9, finance FSC) → schema-per-tenant or database-per-tenant
Standard business → row-level security in PostgreSQL is enough

DV default: every web application has multi-tenant ready architecture (tenant_id in schema, RLS), even if deploy is single-tenant. Convert single → multi in 1-2 sprints instead of 6-month refactoring.

02SCALING AND PERFORMANCE
What happens when the app grows from 100 to 10,000 users?

DV architecture designed for scaling, but bottlenecks appear at specific thresholds:

100-1,000 users: Zero changes in infrastructure, 1 VPS (4-8 GB RAM, 2-4 vCPU) is enough. Cost 100-300 PLN/month.

1,000-10,000 users: Optional read-only PostgreSQL replica for reports (offload primary), Redis cache for hot queries, CDN for static assets. Migration 1-2 sprints. Infrastructure cost 500-1,500 PLN/month.

10,000-100,000 users: horizontal scaling (load balancer + 2-4 app servers), read replicas PostgreSQL, queue processing (BullMQ + Redis), edge caching (Cloudflare). Architecture does not require rewriting - modular structure from day 1 allows to scale elements separately. Cost 2,000-8,000 PLN/month.

100,000+ users: Decision per case - usually requires re-architecting key modules (sharding, microservices where it makes sense, dedicated infrastructure team). DV helps with decisions, but realistically a 100k+ application is usually a different operator than a 10k one.

Database (PostgreSQL) scales to 1M+ users without re-architecting if queries are well indexed and access patterns predictable.

03VENDOR LOCK-IN
Does the DV stack (Next.js + Payload) create vendor lock-in?

Short answer: no, the stack is open-source, the code yours.

Long answer - what specifically protects against lock-in:

Open-source frameworks:

Next.js (MIT license) - used by Netflix, TikTok, Twitch - zero risk"maintainer abandoned".
Payload CMS (MIT license) - platform ownership for CMS, you can host where you want
PostgreSQL (PostgreSQL License) - the most widely used relational database in production, PostgreSQL developer you will find everywhere

No proprietary DV libraries:

Any relationship inpackage.json has a public npm registry
Zero"DV Framework" or"DV Auth Library". which are not available on npm
Configuration deployment in git repo, not in the"DV Cloud Console".

Code in your repo:

GitHub/GitLab Your License
Full git history available
Change of developer/agency = pull request, no"DV cloud code recovery"

A realistic case from our practice: 2 clients switched to in-house team after 18-24 months. Senior developer found his way around the code in 1-2 weeks (Next.js + PostgreSQL = market standard).

The only"lock-in" it'sarchitectural choice - if the application built on Next.js, migration to Vue/Angular requires a frontend rewrite. This is not unique to DV - any agency such"lock-in" creates. DV vs. competition: we choose frameworks with 5+ years of track record, not hype-driven (Stable stack - Block 7 hub/creating-software).

04COMPLIANCE
Is DV building applications under RODO Art. 9 (medicine), FSC (finance) or SOC2?

RODO Article 9 (sensitive data - medical, biometric data): Yes. Baseline DV (encrypted at rest, audit log, RBAC, multi-factor auth) covers core requirements. Discovery (Stage 02) includes RODO gap analysis - what additionally the application needs (e.g. specific retention policies, consent management, DPIA support). Realistically 4-6 RODO Art 9 projects completed, mostly medical and HR.

FSC (financial sector): In part. Baseline of DV web application covers security and audit, butspecific KNF compliance (Recommendations D, M, Z) often requires an external compliance consultant. DV works in the mode ofcollaboration - we build the application, client + KNF compliance company validate. 1-2 fintech projects completed (not full banking, but adjacent - B2B fintech, lending platform).

SOC2 / ISO 27001: Not default. These audits requireoperational processes of the company (continuous monitoring, incident response procedures, employee training, physical security data center) - not just a web application. DV builds the application with technical SOC2-ready controls (audit log, encryption, access control), butcertification is on the client side Or external SOC2 audit of the company.

HIPAA / GDPR US compliance: No. DV works in the Poland / EU / Switzerland markets. For US healthcare compliance we recommend specialized companies.

05MIGRATION FROM LEGACY
Does DV migrate existing applications (PHP legacy, old .NET, custom enterprise)?

Yes, butselectively. Legacy migration is often 60-80% of the budget of a new application, and output only makes sense if the existing systemactually blocks business.

The three migration scenarios we do:

1. UI/frontend migration, backend stays: Most common case. Old PHP/Java/.NET monolith stays, we add a new Next.js frontend communicating with existing backend via API. Time: 8-12 weeks. Risk: low. ROI: fast (client sees modern interface in 2-3 months).

2. step migration (strangler pattern): Old system stays in production, new modules in Next.js + Payload replace legacy modules one by one. Time: 12-24 months. Risk: medium. ROI: spread out, less business disruption.

3. full migration (rebuild): Old system archived (read-only), new system replaces 100% functionality. Time: 16-32 weeks. Risk: high (data migration, user training, rollback strategy). ROI: long-term, but requires business case ROI of 3-5 years.

What we do NOT do:

Migration of a stack that DV does not use in production (Java legacy, Ruby on Rails legacy, custom mainframe)
"Small fixes" In existing legacy code (freelancer scope)
Rebuild without clear business case (custom for custom = fail)

06MOBILE
Can a mobile app be added to the web app later?

Yes - and we often do. DV web application designedAPI-first (Next.js API Routes or a dedicated backend). The mobile app can use the same backend without rewriting the business logic.

Stack mobile: React Native (Expo) - code reuse 60-80% from React/Next.js web application. Shared:

Business logic (TypeScript)
API client
Validation schemas (Zod)
Type definitions

What's specific for mobile:

UI components (mobile patterns, gesture handling)
Push notifications (FCMs/APNs).
Offline mode (if required)
App Store / Play Store deployment

Typical schedule for adding mobile:

6-10 weeks for companion app (reads/edits same data as web app)
12-16 weeks for feature-rich mobile (offline, push, hardware integration)
Price: from PLN 40,000 (companion) to PLN 150,000+ (full custom)

Native (Swift/Kotlin): Only when specific requirements (advanced graphics, performance- critical features, hardware integration like BLE/NFC). In practice DV does React Native in 80%+ of mobile projects. For native contractors, we recommend specialized mobile shops.

07DEPLOYMENT
Does the application need to be on DV servers or can I host at my place?

You can host anywhere you want. DV does not require deployment on your own infrastructure.

Three deployment models:

1. DV-managed VPS (most common, 70% of projects): DV puts VPS on your account, configures, monitoring, backups. You pay for the VPS directly (£80-200/month), DV pays for maintenance separately (£450-1500/month). Root access is yours.

2. client-managed (your infrastructure): Your IT team manages the infrastructure - your own VPS, on-premises server, clustered Kubernetes. DV conveys:

Docker application images
Deployment runbook (step-by-step how to deploy)
Database migration scripts
Configuration template (env vars, secrets)
1-2 knowledge transfer sessions for your team

Mostly for enterprise customers or companies with dedicated DevOps team.

3. Cloud (AWS / GCP / Azure): If the customer already has a cloud ecosystem. DV deploys the app to your cloud account, configure IaC (Terraform/ Pulumi) in git repo, you control billing and access.

Key:

Code in your git repo - regardless of deployment model
Database backups in your cloud storage - zero retention on DV infrastructure
Secrets (API keys, passwords) in your secrets manager - DV does not store customer credentials long-term
Migration path - transition from DV-managed → client-managed always possible (export Docker images + database dump + configuration)

↳ These are 7 questions specific to web applications. General questions about software (build/buy, code ownership, architectural decisions, schedule) - in /our-offer/development-software FAQ block. General questions about working with DV - in /our-offer/development-software FAQ block.

WHEN IT'S WORTH IT - 4 SIGNALS

When a web application pays for itself faster than an off-the-shelf SaaS

Four signals at which custom application is cheaper long-term than trying to squeeze the process into off-the-shelf tools. If you recognize at least two - it's time for an architectural conversation.

01WORKFLOW DOESN'T FIT

Implemented SaaS forces process change instead of support

Specific symptoms:

Custom workflow requires 5+ workarounds in the finished tool
30%+ of functions not used because they don't fit the industry/process
Employees say "simpler to do it in Excel"
Subscriptions to automation tools are on the rise because SaaS-ready lacks key integrations

Custom application: workflow 100% under the company's process, one database, one login, one logic.

02DATA SILOS

Data between systems flows through people, not APIs

Order from marketplace copied manually to accounting. Lead from web form transcribed into CRM. Invoice re-entered into management report. Each operation is 5-15 minutes

risk of error.

Specific symptoms:

3+ systems in the company, data flows through Excel/email/PDF
Backoffice worker spends >20% of time manually copying data
Management reports are late because data needs to be "downloaded"
Each new customer = setup in 4-5 systems separately

Custom application with API integration: real-time synchronization, single source of truth, management reports automatic.

0324/7 SELF-SERVICE

B2B customer expects self-service, we serve by email

Specific symptoms:

The customer asks "do you have a customer panel?"
30%+ of service team's time is repetitive status questions
Returns/complaints generate 5+ emails back and forth
Competitors in the industry have a customer portal, you do not

Custom customer panel: reduce repetitive service by 50-70%, signal market professionalism.

04DECISIONS ON DATA

Management decisions based on late Excel reports

Specific symptoms:

Management reports prepared manually, 1-3 days of work per report
Management asks "how are we doing?" - the answer requires 4 emails to various departments
Data from 3+ systems combined in Excel from scratch each time
Analyst employee overloaded with work, most of which is repetitive

Custom dashboard: real-time data with aggregation, self-service drill-down, alerts for anomalies.

5 SCENARIOS - CHOOSE THE TYPE OF APPLICATION

What specifically are we building

01SAAS

SaaS multi-tenant - from MVP to product

Multi-tenant - Subscriptions - Auth - White-label - Scaling-ready

Stack: Next.js + Payload + PostgreSQL + Stripe. Modular architecture under 3-5 years of iterative development.

Typical project: 8-16 weeks. Price: from PLN 50,000.

02DASHBOARD

Management dashboard with real-time data

BI - Data aggregation - Self-service drill-down - Alerts.

Typical project: 6-12 weeks. Price: from PLN 30,000.

03DASHBOARD

Management dashboard with real-time data

BI - Data aggregation - Self-service drill-down - Alerts.

Typical project: 6-12 weeks. Price: from PLN 30,000.

04CUSTOM CRM

Custom CRM for industries not served by off-the-shelf SaaS

Workflow - Pipeline - Integrations - Industry-specific

Typical project: 8-14 weeks. Price: from PLN 40,000.

Dedicated sub-site in preparation - full description and calculator available soon.

05B2B PORTAL

B2B partner portal for companies of 100-300 people

Partners - Distributors - Agents - White-label - Multi-region.

Typical project: 12-20 weeks. Price: from PLN 80,000.

Dedicated sub-site in preparation - full description and calculator available soon.

↳ The list is not closed. If your case does not fit any scenario (custom marketplace, internal industry platform, employee tool) - describe in the form, we come back with specifics in 24-48h.

WHAT YOU GET - 4 APP-SPECIFIC CONCRETES

What do you get in the price of a DV web application

Four things in the standard of any DV web application - no matter whether SaaS, customer portal, dashboard or custom CRM. Not upsells, not premium packages - the baseline for every project.

UX DESIGNED FOR BUSINESS

UX tailored to real-world workflow, not to design trends

Most web applications have UX"copied from other SaaS". - sidebar nav, dashboard tiles, generic forms. It looks modern, but doesn't match the client company's workflow. Employees click more, not less.

DV works differently:

User research with 3-5 prospective users before design begins (Stage 02 of the process) - what they realistically do, how long, how often
Wireframes tested with users before code (clickable Figma prototype)
Design tokens as CSS variables - visual consistency, but workflow-first
A/B testing of the most important post-launch flow - iterating on data, not guessing

The consequence: employees adopt the application in 1-2 weeks, rather than"we have been learning this for 3 months"..

MONITORING AND ALERTS FOR THE OWNER

You know sooner than the user when something doesn't work

DV standard in every web application:

Sentry - error tracking with full stack trace + session replay (we can see what the user was doing before the error)
Uptime monitoring every 5 minutes - SMS/email alerts when application unavailable >2 minutes
PostHog - product analytics + heatmaps + session recordings
Application Performance Monitoring (response time, DB queries, bottlenecks)
Centralized logging with retention 30 days + alerts for anomalies

Consequence: production error report We to you. Fix typically in 2-4h from incident. Your team does not waste time on"someone says it doesn't work, let's check"..

ARCHITECTURE UNDER 3-5 YEARS OF DEVELOPMENT

An application to which functions are added rather than rewritten

Web applications live for 3-5 years. Architectures designed for this perspective, not for"by launch in 8 weeks".:

Monorepo structure (apps/web, apps/admin, packages/database, packages/ui) - add admin panel or mobile app without rewriting
TypeScript strict mode + Zod validation - contracts between modules checked compile-time (errors caught before deployment)
Database migrations in versioned code - schema evolution without"ALTER TABLE in production".
Multi-tenancy ready (even single-tenant applications) - adding a second client is a configuration, not a refactor
Feature flags + A/B testing infrastructure from day one

Consequence: expansion in 2 years is adding features in 2-4 sprints, not a 6-month refactor of the entire system.

AUTH AND SECURITY AS STANDARD

Security invisible to the user, but present

The web application iscustomer data of your company - A breach means a real legal (RODO), reputational and operational problem. Security baseline as standard:

Authentication: NextAuth/Auth.js or Clerk (multi-factor 2FA, passwordless, SSO optional)
Authorization: role-based access control (RBAC) from the code and database level (defense in depth)
Encrypted at rest (PostgreSQL TDE) + encrypted in transit (TLS 1.3)
OWASP Top 10 covered: SQL injection, XSS, CSRF, broken auth, security misconfiguration
Rate limiting on API endpoints (protection against brute-force and DDoS)
Audit log at the base: who, when, what changed (RODO compliance)

The consequence: the application passes the enterprise client's security audit without panic refactoring. The client's compliance team gets documentation (security checklist) in the handoff.

STACK - 4 LAYERS OF ARCHITECTURE

Technology stack for web applications

01FRONTEND

Next.js + React Server Components + Tailwind

Default frontend for web applications can be: Next.js with App Router and React Server Components. Tailwind CSS + shadcn/ui as component library. React Hook Form + Zod for form validation.

Why this stack:

Next.js App Router - server components reduce JavaScript bundle 30-60% vs pure client React, better LCP/TTI
Tailwind + shadcn/ui - design tokens as CSS variables, component library copy-paste (not node_modules dependency)
Zod validation - single source of truth for types frontend + backend + database

When to withdraw:

Vue / Nuxt - when the customer already has the Vue ecosystem in-house
Astro - for content-heavy applications with minimal interactivity (documentation, information portal)
Angular - enterprise customers with existing Angular database

In 90% of DV web application projects: next.js. The remaining 10% depends on the client.

02BACKEND AND DATABASE

Node.js + TypeScript + PostgreSQL + Prisma

Redis (when needed) or Inngest (serverless background tasks).

Why PostgreSQL:

ACID transactions - zero risk of data loss
JSONB columns - flexibility NoSQL, but in a relational database
Row-level security - multi-tenancy ready without application-layer hacks
Production proven - since 1996, used by Stripe, Reddit, Instagram

Why Prisma:

Type-safe queries - error in queries caught compile-time
Auto-generated migrations with schema diff
Studio (GUI for base) in development, prod-safe in production

Alternatives considered: Drizzle (lighter ORM, edge-deployment), TypeORM (legacy projects), raw SQL (when Prisma constraint).

03AUTH AND MULTI-TENANCY

Auth selected for the application's business model

Auth does not have one"default'a" In web applications - depends on the business model:

Single-tenant (one company, many employees):

NextAuth/Auth.js + email magic links or OAuth (Google, Microsoft) - simplest, zero vendor lock-in
Role-based access control from the base (RBAC) - admin/user/viewer

Multi-tenant SaaS (multiple clients, each with their own workspace):

Clerk or WorkOS - turnkey solutions with multi-org support, SSO, SCIM provisioning
Custom solution on NextAuth (when budget tight or specific requirements)

B2B portal with partners:

SSO (Single Sign-On) from the client system - SAML 2.0 or OAuth
Provisioning automatic from HR of the customer's system

Auth security baseline (each application):

Multi-factor 2FA optional for user
Rate limiting on login endpoints (brute-force protection)
Audit log of each login (RODO compliance)
Sessions with secure HttpOnly cookies

04 INFRASTRUCTURE AND DEPLOYMENT

Infrastructure selected for scale and requirements

The most common DV web application deployment scenarios:

VPS - 80% of projects of Polish clients:

VPS - cost 80-200 PLN/month.
Docker containers, Traefik load balancer, automated SSL
Database PostgreSQL on the same VPS or separate (depending on scale)
Backup daily + retention 30 days

Cloud (AWS / GCP / Azure) - 15% of projects:

When the customer already has a cloud ecosystem (Azure AD, AWS S3, GCP BigQuery)
Serverless architecture (Lambda/Cloud Functions) for event-driven workflows
Managed database (RDS/CloudSQL) - higher cost, but managed by cloud provider

Vercel + Neon/Supabase - 5% of projects:

Global edge network - when an app has users on 3+ continents
Zero DevOps overhead - automated preview deployments per PR
Trade-off: cost increases non-linearly with scale (>200k MAU)

CI/CD any application:

GitHub Actions - automated tests + build + deploy for staging after each merge
Production release manual after code review + acceptance testing

PROCESS - 5 STAGES - 8-16 WEEKS5 steps

From consultation to production - in stages

1-2 weeks - free of charge
ARCHITECTURAL CONSULTATION
First call (60-90 min) + 1-2 follow-up sessions. We map: what business processes the application is supposed to support, what systems exist in the company (CRM, ERP, accounting), what the future functionality roadmap looks like (3-5 years). Build/buy decision - we verify if custom actually makes sense vs ready-made SaaS.
Output: ADR (Architectural Decision Record) v0 + stack decision (frontend / backend / database / auth / infrastructure) with 2-3 alternatives per layer. Each decision has a business case.
2-3 weeks - paid deliverable
DISCOVERY AND SPECIFICATION
Discovery deep-dive specific for web application: user research (3-5 interviews with prospective users), mapping data model (how application organizes data), API integration specification (which external systems, which endpoints, sync frequency, error handling), security model (auth, RBAC, multi-tenancy isolation if SaaS).
Output: 4 heritage tabular documents:
ERD (Entity Relationship Diagram) - database schema.
API contract (OpenAPI 3.0 spec).
Integration maps - which external systems, which operations
User stories + acceptance criteria - testable requirements
For SaaS multi-tenant additionally: tenant isolation strategy (row-level security vs schema-per-tenant vs database-per-tenant).
2-4 weeks
UX/UI + PROTOTYPE TESTED
Wireframes low-fidelity for all key user flows (login, major CRUD operations, reports). Then design tokens and high-fidelity mockups in Figma. Clickable prototype tested with 3-5 usersprior to encoding - We identify 60-70% of UX issues before they become tech debt.
For applications with existing design system of the client (after branding DV or other supplier) - we adapt to existing guidelines instead of creating from scratch.
Clickable Figma prototype + design tokens as CSS variables ready for import in Tailwind config. Zero hand-translation design → code. User testing before coding = reduced re-work by 40-60% in Stage 04.
4-12 weeks - 2-week sprints
DEVELOPMENT + STAGING FROM DAY 1
Longest stage. We work in 2-week sprints with a review after each sprint.Staging environment from the first sprint - Customer tests N+1 feature while working on N+2, feedback loop 1 week instead of 3-month big-bang launch.
MVP scope: core user journey + critical integrations. Secondary features (advanced reports, edge case handling, admin panel deluxe) - after MVP launch based on real user feedback.
Continuous deployment to staging. Code quality in CI/CD: TypeScript strict, ESLint, Vitest unit tests, Playwright E2E for critical paths (login, payment, form submit, API contract). Merge blocked if not passing standards.
Each PR = preview deployment with unique URL - client tests changes isolated from main staging.
Launch + 30 days of stabilization
LAUNCH + MONITORING + STABILIZATION
Production deployment +monitoring stack configured before launch (Sentry, Uptime, PostHog, APM, centralized logging). Handoff session 30-45 min: architecture, deployment process, how to debug production, how to push changes. 30 days of stabilization included (minor fixes, optimizations, bug fixes discovered in production).
After stabilization 3 options:
End of cooperation - zero lock-in, code yours
Technical care subscription
Expanding more features
Handoff documentation heritage tabular:
README + setup instructions (local dev in 30 min)
ADR (Architectural Decision Records) - why this stack
Deployment runbook (how to deploy, how to rollback, how to debug production)
Database schema docs + migrations history
Integration runbooks (how to set up Stripe webhooks, how to rotate API keys, how to debug failed sync)
Playbook monitoring - what alerts mean what, when wakeup-on-call (if critical application)

WEB APPLICATION PROJECT SCHEDULE

8-16 weeks

TOTAL

8-12 weeks

CUSTOMER PORTAL / DASHBOARD

12-20+ weeks

SAAS MULTI-TENANT

↳ Specific schedule after Discovery (Stage 02). SaaS enterprise projects (multi-region, compliance audit, dedicated support) - 20+ weeks. We consult individually.

PRICE LIST - 3 TIERS OF WEB APPLICATION

How much does a web application cost

Each project is priced individually after Discovery. Below 3 tiers of scale typical for DV web applications + calculator for estimating your own project in 4 minutes.

01LITTLE

30 - 60 thousand zlotys.

Time: 6-8 weeks

Single module, 1-2 user roles

1-2 user roles (e.g. admin + user)
5-15 core screens
1-2 API integrations (e.g. Stripe + email)
Single-tenant architecture
Auth standard (NextAuth/Clerk)
Monitoring stack (Sentry + Uptime + PostHog).

02AVERAGE

60 - 120 thousand zlotys.

Time: 8-12 weeks

3-5 roles, integrations, simple multi-tenant

3-5 user roles with RBAC
15-40 core screens
3-5 API integrations (CRM + accounting + payments + email + analytics)
Multi-tenant ready (single-tenant deploy, but architecture for scaling)
Auth with multi-factor + audit log
Background jobs (queue processing, scheduled tasks, email sending)
Full monitoring + alerts

03COMPLETED

120 - 300+ thousand zlotys.

Time: 12-20+ weeks

Multi-tenant SaaS, compliance, advanced features

Multi-tenant SaaS architecture (row-level security or schema-per-tenant)
5+ user roles with complex RBAC + custom permissions
40+ core screens
5-15 API integrations + bidirectional webhooks
Advanced features: white-label, custom fields per tenant, workflow engine, reporting builder
Compliance ready (RODO Art. 9, FSC, audit trail)
Real-time collaboration (WebSocket/SSE).
Subscriptions (Stripe billing) + usage-based pricing

What most strongly determines the price of a web application

Honest Scope:

Hosting and infrastructure in production (VPS 80-200 PLN/month, cloud per usage, dedicated from 500 PLN/month).
Monitoring tools - Sentry, PostHog, Uptime monitoring (Free for small scale, further from ~$25/month per tool)
External services paid per use (Stripe 1.4-2.9% per transaction, Twilio per SMS, OpenAI per token, email delivery from PLN 50/month).
Domains + SSL (~£100/year)
Campaign marketing (App Store Optimization, ads, content)
Content (copywriting, product graphics) - we can identify proven providers

Long-term costs: usually 8-15% per year of the project value for hosting + technical care + minor extensions.

What our forks don't cover

Heritage scope - what's outside our area of work:

Lean MVP for 10-20 thousand zlotys. - At this budget, a throwaway code is usually created. Better options: no-code (Bubble, FlutterFlow) as first validation, senior freelancer with portfolio, startup gas pedal with built-in tech support.
Custom enterprise with budgets of £1mil. - multi-region failover, custom hardware, dedicated support team 24/7 - fit larger consulting firms (100+ person software house).
Migration of legacy enterprise systems (mainframe, JEE legacy, custom ERPs 15+ years old) - DV does not work with legacy stack.
Single fixes in another agency's existing application - We work on a project basis (Discovery → Deployment → Long-term care), not hourly for minutiae.

If your project falls into the above categories - we say so directly instead of taking a project that we will not do optimally.

What is NOT included in the project price

Honest Scope:

Hosting and infrastructure in production (VPS 80-200 PLN/month, cloud per usage, dedicated from 500 PLN/month).
Monitoring tools - Sentry, PostHog, Uptime monitoring (Free for small scale, further from ~$25/month per tool)
External services paid per use (Stripe 1.4-2.9% per transaction, Twilio per SMS, OpenAI per token, email delivery from PLN 50/month).
Domains + SSL (~£100/year)
Campaign marketing (App Store Optimization, ads, content)
Content (copywriting, product graphics) - we can identify proven providers

Long-term costs: usually 8-15% per year of the project value for hosting + technical care + minor extensions.

Count the cost of a web application Describe your project

WHAT MAKES US DIFFERENT - 3 APP-SPECIFIC THINGS

3 things we do differently in web applications

MULTI-TENANCY READY

Multi-tenant architecture even for single-tenant projects

DV works differently. Any web application designedmulti-tenant ready from the first sprint, even if the deploy is single-tenant:

Database schema withtenant_id in each table (even for single-tenant - tenant_id = '1')
Row-level security (RLS) in PostgreSQL - isolating data at the database level, not the application level
Tenant routing (subdomain or path-based) - ready for activation in 1-2 sprints
Per-tenant config (logo, primary color, feature flags, pricing plans) - schema ready, UI panel added on-demand
Middleware auth with tenant context - request always knows its tenant

PRODUCTION-READY

Staging from sprint 1, pre-launch monitoring, zero "big bang"

DV works differently:

Staging environment from sprint 1 - link current always, client is testing feature N+1 while working on N+2
Continuous deployment - any merge to main = auto-deploy to staging in 5-10 mins
Preview deployments per PR - each pull request = unique URL with isolated deployment
Code quality in CI/CD - TypeScript strict, ESLint, Vitest unit tests, Playwright E2E for critical paths. Mergeblocked if it doesn't pass standards
Monitoring stack configured before launch - Sentry, Uptime, PostHog, APM, centralized logging. Not"we will add after the launch"., only"works before the first user"

CODE OWNERSHIP

Code in your repo from the first sprint, zero proprietary DV libraries

DV works differently from the first sprint:

Full source code in your GitHub/GitLab - From first commit, not after launch, not after last invoice
Your license, your intellectual property - zero clauses"DV retains the rights to the code". contract
Open-source / standard frameworks only - Next.js, Payload, PostgreSQL, Stripe - zero proprietary DV libraries that are not on npm
Zero hidden integrations - each third-party integration has a trail audit in the code + documentation on how to configure it under the client's own account
Handoff documentation included - README + ADR + deployment runbook + database schema docs + integration runbooks (6 heritage tabular documents)

FOR WHOM DV - AND FOR WHOM NOT

Who we build web applications for - and who we don't

Company of 30-150 people, custom system as a must-have

What we give you: dedicated application under your workflow, multi-tenant ready architecture from day 1, monitoring and long-term technical care (450-1500 PLN/month).

Ready-made SaaS covers 70%+ of your needs

Custom web application islong-term investment (PLN 50-300k + 5-15% per year maintenance). If a ready-made SaaS covers most of your needs - buy, don't build.

The most common situations when we say "buy SaaS ready":

CRM for 10-30 people - Pipedrive, HubSpot Free, Bitrix24 starter
Project management - turnkey project management SaaS, industry SaaS (e.g., for architects, lawyers).
Mailing automation - Brevo, Mailchimp, ActiveCampaign
Booking/reservation - Calendly, Cal.com, industry SaaS
Helpdesk/ticketing - turnkey solutions from ~£200/month.

In a 20-minute architectural consultation, DV will walk you through a build/buy decision. 15-20% of the conversations end with the recommendation "don't build, buy ready-made SaaS X".

Startup post-MVP, refactor or rebuild under scale

What you get with DV: production-ready stack from day one, multi-tenant architecture, monitoring + alerts, due-diligence documentation in handoff.

Lean MVP for 10-20 thousand zloty / proof-of-concept in a week.

Recommended alternatives:

No-code platform - Bubble (strongest for web apps), FlutterFlow (mobile + web), Glide (simple internal tools)
Freelancer senior With a portfolio of similar projects (Useme, Bulldogjob, Polish tech communities)
Startup gas pedal with built-in tech support (PARP Startup Platforms, Startup Booster Poland - funding)

Come back to us when the MVP validates the business hypothesis and you need a production-ready system under scale (100+ users, multi-tenant, SLA, compliance).

Company of 100-300 people, outsourcing a module of a larger system

Enterprise 500+ employees, budget 1+ mil, multi-region

Fit larger consulting firms (100+ people software house Polish/international) or direct hire enterprise dev team.

FAQ - 7 QUESTIONS ABOUT WEB APPLICATIONS

Questions we get during the consultation phase

01MULTI-TENANCY
When should an application be multi-tenant, and when is single-tenant enough?

Three decision criteria:

Will the application support >1 customer/team/brand in the future 3-5 years?

Yes (even hypothetically) → multi-tenant ready from day 1
Not sure → multi-tenant ready architecture, but single-tenant deploy (DV default)
Definitely not → single-tenant (rare in practice)

Is the app ever going to be white-label for resellers?

Yes → multi-tenant required, custom branding per tenant
No → single-tenant ok

Compliance requirement to isolate data between customers?

Yes (medicine RODO Art. 9, finance FSC) → schema-per-tenant or database-per-tenant
Standard business → row-level security in PostgreSQL is enough

02SCALING AND PERFORMANCE
What happens when the app grows from 100 to 10,000 users?

DV architecture designed for scaling, but bottlenecks appear at specific thresholds:

100-1,000 users: Zero changes in infrastructure, 1 VPS (4-8 GB RAM, 2-4 vCPU) is enough. Cost 100-300 PLN/month.

Database (PostgreSQL) scales to 1M+ users without re-architecting if queries are well indexed and access patterns predictable.

03VENDOR LOCK-IN
Does the DV stack (Next.js + Payload) create vendor lock-in?

Short answer: no, the stack is open-source, the code yours.

Long answer - what specifically protects against lock-in:

Open-source frameworks:

Next.js (MIT license) - used by Netflix, TikTok, Twitch - zero risk"maintainer abandoned".
Payload CMS (MIT license) - platform ownership for CMS, you can host where you want
PostgreSQL (PostgreSQL License) - the most widely used relational database in production, PostgreSQL developer you will find everywhere

No proprietary DV libraries:

Any relationship inpackage.json has a public npm registry
Zero"DV Framework" or"DV Auth Library". which are not available on npm
Configuration deployment in git repo, not in the"DV Cloud Console".

Code in your repo:

GitHub/GitLab Your License
Full git history available
Change of developer/agency = pull request, no"DV cloud code recovery"

A realistic case from our practice: 2 clients switched to in-house team after 18-24 months. Senior developer found his way around the code in 1-2 weeks (Next.js + PostgreSQL = market standard).

04COMPLIANCE
Is DV building applications under RODO Art. 9 (medicine), FSC (finance) or SOC2?

HIPAA / GDPR US compliance: No. DV works in the Poland / EU / Switzerland markets. For US healthcare compliance we recommend specialized companies.

05MIGRATION FROM LEGACY
Does DV migrate existing applications (PHP legacy, old .NET, custom enterprise)?

Yes, butselectively. Legacy migration is often 60-80% of the budget of a new application, and output only makes sense if the existing systemactually blocks business.

The three migration scenarios we do:

What we do NOT do:

Migration of a stack that DV does not use in production (Java legacy, Ruby on Rails legacy, custom mainframe)
"Small fixes" In existing legacy code (freelancer scope)
Rebuild without clear business case (custom for custom = fail)

06MOBILE
Can a mobile app be added to the web app later?

Yes - and we often do. DV web application designedAPI-first (Next.js API Routes or a dedicated backend). The mobile app can use the same backend without rewriting the business logic.

Stack mobile: React Native (Expo) - code reuse 60-80% from React/Next.js web application. Shared:

Business logic (TypeScript)
API client
Validation schemas (Zod)
Type definitions

What's specific for mobile:

UI components (mobile patterns, gesture handling)
Push notifications (FCMs/APNs).
Offline mode (if required)
App Store / Play Store deployment

Typical schedule for adding mobile:

6-10 weeks for companion app (reads/edits same data as web app)
12-16 weeks for feature-rich mobile (offline, push, hardware integration)
Price: from PLN 40,000 (companion) to PLN 150,000+ (full custom)

07DEPLOYMENT
Does the application need to be on DV servers or can I host at my place?

You can host anywhere you want. DV does not require deployment on your own infrastructure.

Three deployment models:

2. client-managed (your infrastructure): Your IT team manages the infrastructure - your own VPS, on-premises server, clustered Kubernetes. DV conveys:

Docker application images
Deployment runbook (step-by-step how to deploy)
Database migration scripts
Configuration template (env vars, secrets)
1-2 knowledge transfer sessions for your team

Mostly for enterprise customers or companies with dedicated DevOps team.

Key:

Code in your git repo - regardless of deployment model
Database backups in your cloud storage - zero retention on DV infrastructure
Secrets (API keys, passwords) in your secrets manager - DV does not store customer credentials long-term
Migration path - transition from DV-managed → client-managed always possible (export Docker images + database dump + configuration)

Web application for a company that has outgrown its sheets

Web application for a company that has outgrown its sheets

When a web application pays for itself faster than an off-the-shelf SaaS

Implemented SaaS forces process change instead of support

Data between systems flows through people, not APIs

B2B customer expects self-service, we serve by email

Management decisions based on late Excel reports

What specifically are we building

SaaS multi-tenant - from MVP to product

Multi-tenant - Subscriptions - Auth - White-label - Scaling-ready

Management dashboard with real-time data

BI - Data aggregation - Self-service drill-down - Alerts.

Management dashboard with real-time data

BI - Data aggregation - Self-service drill-down - Alerts.

Custom CRM for industries not served by off-the-shelf SaaS

Workflow - Pipeline - Integrations - Industry-specific

B2B partner portal for companies of 100-300 people

Partners - Distributors - Agents - White-label - Multi-region.

What do you get in the price of a DV web application

UX tailored to real-world workflow, not to design trends

You know sooner than the user when something doesn't work

An application to which functions are added rather than rewritten

Security invisible to the user, but present

Technology stack for web applications

Next.js + React Server Components + Tailwind

Node.js + TypeScript + PostgreSQL + Prisma

Auth selected for the application's business model

Infrastructure selected for scale and requirements

From consultation to production - in stages

How much does a web application cost

30 - 60 thousand zlotys.

Single module, 1-2 user roles

60 - 120 thousand zlotys.

3-5 roles, integrations, simple multi-tenant

120 - 300+ thousand zlotys.

Multi-tenant SaaS, compliance, advanced features

What most strongly determines the price of a web application

What our forks don't cover

What is NOT included in the project price

3 things we do differently in web applications

Multi-tenant architecture even for single-tenant projects

Staging from sprint 1, pre-launch monitoring, zero "big bang"

Code in your repo from the first sprint, zero proprietary DV libraries

Who we build web applications for - and who we don't

Company of 30-150 people, custom system as a must-have

Ready-made SaaS covers 70%+ of your needs

Startup post-MVP, refactor or rebuild under scale

Lean MVP for 10-20 thousand zloty / proof-of-concept in a week.

Company of 100-300 people, outsourcing a module of a larger system

Enterprise 500+ employees, budget 1+ mil, multi-region

Questions we get during the consultation phase

01|MULTI-TENANCYWhen should an application be multi-tenant, and when is single-tenant enough?

02|SCALING AND PERFORMANCEWhat happens when the app grows from 100 to 10,000 users?

03|VENDOR LOCK-INDoes the DV stack (Next.js + Payload) create vendor lock-in?

04|COMPLIANCEIs DV building applications under RODO Art. 9 (medicine), FSC (finance) or SOC2?

05|MIGRATION FROM LEGACYDoes DV migrate existing applications (PHP legacy, old .NET, custom enterprise)?

06|MOBILECan a mobile app be added to the web app later?

07|DEPLOYMENTDoes the application need to be on DV servers or can I host at my place?

When a web application pays for itself faster than an off-the-shelf SaaS

Implemented SaaS forces process change instead of support

Data between systems flows through people, not APIs

B2B customer expects self-service, we serve by email

Management decisions based on late Excel reports

What specifically are we building

SaaS multi-tenant - from MVP to product

Multi-tenant - Subscriptions - Auth - White-label - Scaling-ready

Management dashboard with real-time data

BI - Data aggregation - Self-service drill-down - Alerts.

Management dashboard with real-time data

BI - Data aggregation - Self-service drill-down - Alerts.

Custom CRM for industries not served by off-the-shelf SaaS

Workflow - Pipeline - Integrations - Industry-specific

B2B partner portal for companies of 100-300 people

Partners - Distributors - Agents - White-label - Multi-region.

What do you get in the price of a DV web application

UX tailored to real-world workflow, not to design trends

You know sooner than the user when something doesn't work

An application to which functions are added rather than rewritten

Security invisible to the user, but present

Technology stack for web applications

01MULTI-TENANCY
When should an application be multi-tenant, and when is single-tenant enough?

02SCALING AND PERFORMANCE
What happens when the app grows from 100 to 10,000 users?

03VENDOR LOCK-IN
Does the DV stack (Next.js + Payload) create vendor lock-in?

04COMPLIANCE
Is DV building applications under RODO Art. 9 (medicine), FSC (finance) or SOC2?

05MIGRATION FROM LEGACY
Does DV migrate existing applications (PHP legacy, old .NET, custom enterprise)?

06MOBILE
Can a mobile app be added to the web app later?

07DEPLOYMENT
Does the application need to be on DV servers or can I host at my place?

01MULTI-TENANCY
When should an application be multi-tenant, and when is single-tenant enough?

02SCALING AND PERFORMANCE
What happens when the app grows from 100 to 10,000 users?

03VENDOR LOCK-IN
Does the DV stack (Next.js + Payload) create vendor lock-in?

04COMPLIANCE
Is DV building applications under RODO Art. 9 (medicine), FSC (finance) or SOC2?

05MIGRATION FROM LEGACY
Does DV migrate existing applications (PHP legacy, old .NET, custom enterprise)?

06MOBILE
Can a mobile app be added to the web app later?

07DEPLOYMENT
Does the application need to be on DV servers or can I host at my place?