We build web and mobile applications, ERP/CRM systems, e-commerce platforms, and custom AI agents. We also help with integrations (Stripe, GA4, Plausible, custom APIs) and long-term maintenance.
Short projects 4-8 weeks, mid-size 3-6 months, long-term partnerships 1+ year. We often continue on a retainer model.
Both models. Fixed price for clearly defined scopes, T&M for exploratory phases and long-term partnerships. We always recommend the model that is fair for both sides.
Yes. We have a dedicated "startup track" with a fast MVP cycle (4-6 weeks) that lets you validate your product on the market before committing to a full investment. For smaller companies we offer packaged pricing.
A dedicated project manager + a shared Slack/Teams channel + a weekly async update in Notion/Linear. You always know what is being worked on, what has moved forward, and what the next milestone is.
Frontend: Next.js / React / TypeScript. Backend: Node.js / NestJS, Python for AI/ML. Databases: PostgreSQL + Prisma, Qdrant for vectors. Infra: Coolify on VPS, Docker, GitHub Actions.
Yes. Two-week iterations, daily async standups, and a demo at the end of each iteration. The client has continuous access to the roadmap and backlog.
As a standard. A forensic-grade audit (OWASP top 10, dead code, performance) is part of every major release. We also offer standalone security audits for existing projects.
Of course. All code is your property from the very first commit. We hand over a clean repository, documentation, onboarding for your team, and test coverage of the key areas.
Typically GitHub Actions with automated linting, typechecking, tests, and builds. Deployment via Coolify (a preview environment for every PR, production deploys from `main`). One-click rollback in the UI.
A system where an AI agent (e.g. built on a RAG architecture) works directly with your company data — helping with lead scoring, churn prediction, and report generation.
Claude (Anthropic) and GPT (OpenAI) for production deployments, local Llama models for sensitive use cases. The architecture is provider-agnostic.
Our architecture uses a thin adapter layer, so swapping the model (Claude 4 → GPT-5 → your own) is a matter of configuration, not refactoring.
Always within your own infrastructure (on-prem or your cloud tenant). For LLM queries we send only the necessary context — sensitive data never goes to public APIs. Under strict regulations (GDPR, HIPAA) we deploy local self-hosted models.
Repetitive decision-making with well-defined inputs/outputs (ticket classification, report generation, lead scoring, content drafts, form validation). Less suitable are tasks requiring critical business decisions without an audit trail.
Monthly plans (Start to Enterprise) are invoiced at the start of each period. One-off custom development: a 50% deposit at signing, 30% at the project midpoint and 20% at handover; large projects use individually agreed milestones. Prices exclude VAT.
Monthly plans already include hosting and operations. For one-off custom development, hosting is not included in the price — we recommend our managed hosting on Coolify/Hetzner starting at CZK 1,500/month.
We invoice by bank transfer as a standard. For international clients we offer Stripe (with an additional 2.9% fee).
Bug fixes during the warranty period (90 days) are free. After the warranty we offer retainer packages (5/10/20 hours per month) with priority SLA response, monitoring, and regular dependency updates.
We refer you to partners specializing in grant consulting (TACR, OP TAK, regional funds). For grant-funded projects we prepare the technical specification and budget in the format required by the grant providers.