Do we own the code?
Yes, fully. You receive the entire source code including documentation and can develop it further, run it yourself or have it continued by third parties without restriction. There is no licensing lock-in to us.
What does “production-ready” mean to us?
Production-ready means, for us: the backend is not just a local prototype but deployable and operable — with clean, documented APIs, security fundamentals (secret handling, input validation, rate limiting, TLS), GDPR fundamentals (data export and deletion, PII-safe logs, EU hosting), traceable AI costs and correct behaviour under failure (no half-written record on abort, atomic transactions). Plus tests for the core paths, handover documentation and the full source code. Honestly delimited: production-ready means running solidly and securely — not automatically scaled for high load or equipped with full operational monitoring. The latter is delivered by the production building blocks (add-ons).
Is the software GDPR-compliant?
We deliver the technical prerequisites for GDPR-compliant operation — and honestly delimit what of that is technology and what is not. Included are: data export, full deletion (Art. 17, implemented end-to-end, with no residual personal data left behind), data minimisation, PII-safe logs, EU hosting and transport encryption. What cannot be produced by code alone — and is therefore explicitly not part of our promise — is overall legal compliance: it arises only from contract, process and technology together, such as a data processing agreement, a record of processing activities and a documented deletion concept. These contractual and organisational building blocks rest with you or your legal advisors; we provide the technical foundation for them.
Which package fits — API Starter or Backend + Frontend?
The API Starter is the backend on its own: you or your team build the frontend and use our APIs. Backend + Frontend is the complete product — the same backend plus a finished interface with login, roles and permissions and the screens of your product (e.g. CRM, chatbot, internal tool). The price difference reflects exactly that: with Backend + Frontend, an entire second application is added — the frontend including user management and a permission system, not just “a bit more backend”. If you have your own frontend team, the API Starter is enough; if you want to ship ready to launch, Backend + Frontend is the way.
What does the frontend in “Backend + Frontend” cover?
A finished, functional web frontend built on proven UI components (a clean, responsive interface) — tailored to your product, from a CRM through a chatbot to an internal tool. The stated price is an entry point for a typical core scope: login and account plus the central functional screens of your product. We agree scope and complexity together before the project starts and fix a binding price on that basis — more screens or a heavyweight core feature raise it accordingly, but it is fixed before we begin, with no hourly surprises during the project. Not included are extensive custom design work or branding from scratch and native mobile apps.
What is included in the base — and what is an add-on?
The base delivers a correctly and securely running backend: clean APIs, an encapsulated AI layer, security and GDPR fundamentals, traceable AI costs and correct behaviour under failure — including structured logging with correlation IDs and health checks. The production building blocks (add-ons) add the operational tooling on top: metrics and dashboards (Grafana) for operational insight, actively enforced usage limits, tested failure paths when the AI provider goes down or degrades, and database-enforced, multiply-secured tenant separation. In short: the base works; the add-ons make operation visible, actively cap costs and provably safeguard edge cases.
What does ongoing operation cost?
You pay hosting and AI usage directly with the respective provider — transparent and in your hands. AI costs are traceably captured (logged) in every package. To actively cap consumption, add the AI Usage Limit add-on — a per-user message limit plus a global request limit. Typical operating costs for small applications are in the low three-digit range per month and only rise with real usage.
Can we integrate any package into our own infrastructure?
Yes. You receive the full source code and a containerised, portable application with clean configuration — it is not tied to our environment and runs in your own infrastructure (cloud, Kubernetes, on-premise). The fixed price includes deployment to one standard environment; with the code and documentation your team can roll out anywhere. If you want us to actively handle the integration into your infrastructure — connecting to your cloud, networking and secrets management, your CI/CD pipelines, SSO/IAM, monitoring stack or on-premise requirements — that is customer-specific effort that cannot be fixed-priced upfront. We bill that part transparently at a day rate, with the scope agreed in advance. If you deploy yourself, there are no additional costs.
What caps AI costs — and what do the AI Usage Limit and AI Budget Cap do?
In the base, AI cost is captured per request and is traceable — it is logged, but not actively limited. Enforcement comes as an add-on: the AI Usage Limit caps consumption by request count — a per-user message limit and a global daily limit across all requests. Both apply automatically — once a limit is reached, further requests are rejected instead of letting consumption and costs run on unbounded. This keeps AI spend predictably capped. The AI Budget Cap works one level deeper and caps not the number of requests but the actual cost: since not every request costs the same — a long answer or a larger model weighs more heavily — it sums running consumption in US dollars and stops further AI calls once the set budget is reached (with a small, concurrency-related overshoot). Its added value: protection directly at the cost level, even with few but expensive requests. Both add-ons can be combined.
What do observability and resilience mean?
Observability means you can see how the system is running — through metrics and Grafana dashboards (requests, latencies, AI consumption). That is delivered by the “Observability (Metrics & Dashboards)” add-on. Resilience means the system behaves in a controlled way when the AI provider fails, slows down or returns faulty responses — with tested failure paths and automatic handling instead of errors propagating through. That comes with the “Observability + Resilience” bundle, including a test suite that deliberately provokes and safeguards exactly these failure modes. Basic runtime insight — structured logs with correlation IDs and health checks — is already on board in every package; the add-ons add the visibility through dashboards and the safeguarded failure handling on top.
How does modernising existing software work?
We always start with a paid discovery/audit. In it we examine your existing architecture, identify integration points and risks and decide honestly whether a solid fixed price is possible. The discovery fee is credited against the follow-up engagement. So there is no fixed price out of thin air — and if an integration cannot be done responsibly, we say so before costs arise. We work in the JVM/Spring/Kotlin environment.
What does the asterisk (*) on the discovery price mean?
The discovery fee (from €2,500) is creditable: if it leads to a follow-up engagement, it is credited in full against the integration price. It only remains a standalone cost if no integration follows the discovery — for instance because we honestly advise against it.
Do you offer maintenance?
Optional — from pure self-operation (you take over the code including documentation) through occasional support to a continuously managed solution. You decide how much responsibility stays with you.
What does an extension after project completion cost?
Extensions beyond the agreed fixed-price scope are billed transparently at a day rate. We align scope and effort in advance so there are no surprises.
Are there introductory rates?
For a limited number of first projects, yes. In return we ask for a nameable reference after successful completion. Bring it up in the initial call.
