Direct answers, no filter.

An ethical software house designs digital products starting from human impact, not from extractive engagement. In practice this means refusing dark patterns, optimising for user well-being rather than time spent in the app, choosing sober stacks that consume less, and being transparent on costs, timelines and technical limits. At BeVibes.Tech we apply this approach to every project, both for external clients and our internal products like BeVibesApp.

Human-centered design means designing from the real needs of the people who will use the product, not from team preferences or market trends. It's done with user research, interviews, and prototypes tested before code. For us it's a method, not a label: every BeVibes interface comes from at least one feedback iteration with real users in the target sector.

GEO is the optimisation of a website to be cited by generative AI engines (ChatGPT, Perplexity, Claude, Gemini, Copilot) as a source of answers. It's the evolution of classic SEO: where SEO aims to rank a site at the top of Google results, GEO aims to make a brand appear inside AI answers. It requires structured content, rich JSON-LD, clear FAQs, author authority, and answer-ready language (first sentence = direct answer). BeVibes.Tech applies GEO to its own site and offers it as a service to clients.

Sustainable technology means writing and running software that consumes the least resources possible, in measurable ways. It involves concrete choices: no redundant API calls, aggressive caching, lightweight JavaScript bundles, optimised images (WebP/AVIF), edge functions instead of always-on servers, hosting on providers with verified green energy. For us it's also ethical: every wasted request is wasted electricity, and technical efficiency is the first form of environmental respect for software builders.

Our core stack is: Next.js + TypeScript + Tailwind CSS for web frontend, React Native + Expo for mobile, Supabase + PostgreSQL for backend and database, OpenAI / Anthropic / Vercel AI SDK for AI integrations, Vercel for hosting. We add Stripe for payments, Resend for email, Sentry and PostHog for monitoring. We don't sell the stack as dogma: we pick it per project. If a client is on a different infrastructure and it makes sense, we adapt.

BeVibes.Tech is registered at Via del Braldo 86/a, 47121 Forlì (FC), Italy, and operates from a studio at Viale Giacomo Matteotti 54/D, 47122 Forlì. We work with clients across Italy and Europe in hybrid mode: in-person meetings when they add value, remote-first for daily operations. We regularly collaborate with H2Bit in Bologna on shared management software projects.

BeVibes.Tech was founded by Nicola Decesari (Founder & CEO) and Lorena Garavini (Co-founder). The operating team includes Alessia Primerano (Chief Financial Officer) and Massimo Tamburini (Fullstack Developer). On shared management software projects (Hostipando, Arti Motorie, e-commerce) we work in tight partnership with Matteo Mascellani, CTO of H2Bit in Bologna. You'll find everyone's LinkedIn profiles on the /team page.

Yes. Our most solid structural collaboration is with H2Bit (Bologna), with whom we co-develop vertical management software: Hostipando for B&Bs, Arti Motorie for physiotherapy, and an upcoming e-commerce project. We're open to partnerships with other studios when a project requires complementary skills. We believe small and mid-size software houses that specialise and integrate serve the client better than a monolithic structure.

Nicola Decesari, founder and CEO, holds the Google AI Professional Certificate, earned on 2 September 2026 by completing all 8 courses in the programme: AI Fundamentals, AI for Brainstorming and Planning, AI for Research and Insights, AI for Writing and Communicating, AI for Content Creation, AI for Data Analysis, AI for App Building, AI for App Deployment. It is a Google Career Certificate delivered through Coursera, issued to him as an individual rather than to the company, and each course also carries its own badge with a public verification page on Credly. At company level, BeVibes S.r.l. is a Partner 24 ORE Business Owner 2026. We do not claim anything a third party cannot verify.

A software house builds software to order: it understands a company's process, designs the tool that runs it, and maintains it over time. The difference from a web agency is the kind of work, not the size: an agency delivers sites and communication, a software house writes applications that handle data, roles and business rules. We do both, but the second is where we come from.

Yes, we accept European clients and occasionally extra-European ones. We work in Italian and English, manage international contracts and billing under EU rules, and adapt schedules for remote team overlap. We prefer SMEs and tech startups with an identifiable product owner: our processes need interlocutors who decide, not committees who approve. Primary active markets: Italy, Germany, France, Spain, Netherlands.

We've worked in: hospitality and food (Oasi StreetFood: mobile-first digital menu), wellness and sport (Asanafy: yoga + AI posture, BeVibesApp: smile-to-earn), accommodation (Hostipando: B&B management), healthcare and rehabilitation (Arti Motorie: whitelabel physiotherapy management), e-commerce retail (in development with H2Bit). Our edge isn't the sector but the method: human-centered design + sober stack + transparency. It works wherever there's a real operational problem to solve with software.

The subscription, nearly always, as long as it covers the work. Custom becomes worth it when the manual work left outside costs more than the price difference, or when the process that sets you apart is exactly the one no standard product runs. If you are in the first case we will say so, even when it means not selling you anything: we distribute Presenzialista, an attendance system from 19 euro a month, precisely for the cases where building from scratch would be a waste.

You pick one step, measure what it costs today, and automate that while leaving the rest alone. If it holds for a few weeks on real data, you move to the next one. Rewriting the whole process around AI is the fastest way to never ship it. Where a mistake is cheap the automation proceeds alone; where it is not, it proposes and a person confirms, and every step is recorded.

Yes, with four separate programmes for people doing different jobs: AI for everyday work, for data analysis, for prototyping internal tools, and half a day on processes for decision makers. They run live online, one to one, as a workshop at your offices or as a company programme spread over time. At the end we issue a BeVibes.Tech certificate of attendance, which is not a certification recognised by a third-party body, and we will not word it as if it were.

A serious custom app starts at €25,000-40,000 for a functional MVP on a single platform with 3-5 core features, rises to €60,000-120,000 for multi-platform products with dedicated backend and third-party integrations, exceeds €150,000 when AI, computer vision, marketplace or enterprise flows are involved. Below €20,000 you enter the territory of no-code or adapted templates: legitimate but with structural limits. BeVibes.Tech always provides a binding estimate at proposal stage, with no surprises mid-project.

For a single-platform MVP with well-defined core features: 6-10 weeks from kick-off to beta release. For a multi-platform product with backend, authentication, payments and analytics: 3-5 months. Complex AI or management projects: 5-9 months including testing and model training. Real timelines almost always depend on client decision speed, not code-writing speed. A product owner who replies in 24h halves the calendar compared to one who replies in 7 days.

Yes. After release we offer three options: monitoring & hotfix (rolling, hourly cost under SLA), monthly retainer with prepaid hours (feature evolution + maintenance, discounted vs spot hourly), or full handover if the client has an internal team. We don't lock anyone into yearly contracts: a well-built product shouldn't need continuous assistance, and if it does it usually means something in the architecture needs revisiting.

Through the form on /contact or by writing directly to info@bevibes.tech. The more detail you share (objective, target users, features already in mind, timing/budget constraints), the more precise the first answer. Within 24 business hours you receive a first non-binding read: we tell you if the project fits us, a realistic cost range, and the next steps. No multi-step forms, no AI assistant pre-filtering: a team member always replies to you.

A vertical management system with an admin area and a user area typically starts in the tens of thousands; a first version narrowed to one flow costs considerably less. The figure depends on how much it has to do, and a first version always does less than the list you start with: that is why the first step is writing the scope down, which produces an estimate with a range rather than a single figure.

For a modern tech startup that wants performance, code control and uncompromised scalability: Next.js. For a startup with a non-technical team publishing heavy editorial content and needing a visual CMS: WordPress (or headless WordPress + Next.js on top). Next.js wins on speed, SEO/GEO, security, ability to integrate AI and dynamic products. WordPress wins on time-to-market for content-heavy sites managed by editors. The right question is: is it a product or a site? If product, Next.js. If content site, WordPress can work.

Supabase uses PostgreSQL (standard relational database) and offers auth, storage, edge functions, real-time. Firebase uses Firestore (NoSQL document-based) with the same range of services. Practical differences: Supabase gives you standard SQL, easy complex queries, full data exportability (it's just Postgres), possible self-hosting, more transparent pricing. Firebase has tighter Google Cloud integration, more mature real-time for chat scale-out, but Firestore sometimes forces awkward data models. For most projects we run, Supabase. For Google multi-cloud cases or massive real-time, Firebase can make sense.

Start from a concrete problem, not the technology. We identify where AI actually reduces user friction or operational time: semantic search, text generation, classification, recommendations, voice synthesis, computer vision, chatbots on proprietary knowledge bases. Then we pick the approach: external API (OpenAI, Anthropic, Mistral) for development speed, self-hosted model for privacy or cost control, on-device for zero latency and zero runtime cost. For most business cases, an API call + aggressive caching + good prompt engineering delivers 80% of the value with 20% of the work.

React Native is a framework that lets you write a mobile app in JavaScript/TypeScript and compile it into native iOS and Android apps, sharing most of the code between platforms. Main benefit: one team, one codebase, two native apps shippable to stores. Pick it when: you need cross-platform mobile with a rational budget, want iteration speed, easy integration with a shared web backend. Don't pick it when: you need extreme 3D performance, deep access to ultra-native features (e.g. CarPlay, professional audio), or the app will ever be iOS-only.

No. We use the OpenAI and Anthropic APIs in modes that do not use submitted data for training. Where data is sensitive, though, a contractual guarantee is not the point, and the answer is different: an open model running on a machine of yours, or processing on the device of whoever uses it, as in Asanafy where the computer vision runs in the browser. If the data never crosses the network, there is nothing left to guarantee.

No-code (Bubble, Webflow, Glide) is excellent for: fast prototypes to validate, MVPs with standard business logic, internal dashboards, experimental landings. It stops working when: logic gets complex, you need custom integrations the tool didn't anticipate, performance matters, you want real code ownership. Custom (Next.js + Supabase) is the right jump once you've validated the problem, know the flows, and want a product that scales and belongs to you. The healthy sequence is: no-code to validate → custom to scale. Skipping the first step costs more time, not less.

Three short phases. 1. Discovery call (45 min): we understand the goal, user, constraints and already say honestly if the project fits us. 2. Written proposal (3-5 business days): document with scope, phases, timing, binding cost, allocated team and known risks. 3. Operational kick-off (1 week): contract signed, shared workspace (Linear or Notion), repo, first design check, first screens. From there, weekly staging releases visible to the client. No fuzzy initial phase, no endless discovery: we want whoever pays to see a real product within 30 days.

Yours. Repository, infrastructure and credentials are in your name from day one, not handed over at the end: if you ever change supplier there is nothing to retrieve, it is already where it belongs. The same goes for hosting accounts, where we work inside them rather than sitting between you and the service.

Yes, and it is a good share of the work: projects inherited from another supplier, applications nobody has checked in months, systems that do not talk to each other. It always starts with a read of the code and the infrastructure, to know what is there before promising anything, and it produces a map of where everything lives and who has access to what. If the project is in a state where rebuilding beats repairing, we say so with the numbers in hand.