Technologies we use to build software, apps and AI
Digitup Studio is a technical studio in Madrid that builds software, apps and AI on a well-known, maintainable stack: Angular and Ionic, Node.js and Python, PostgreSQL or Firebase, and models from OpenAI, Anthropic or Gemini. We choose each piece to fit the project, not out of habit.
In short
- Frontend and apps: Angular, Ionic and TypeScript, with one codebase for iOS and Android.
- Backend and data: Node.js with NestJS, Python with FastAPI or Firebase; PostgreSQL, MongoDB, Redis or Firestore.
- AI: OpenAI, Anthropic, Gemini, DeepSeek and Perplexity, with RAG, embeddings and agents.
- Production: VPS or cloud, Docker, CI/CD and monitoring.
- Existing product: we review it on the first call.
Stack by layer
Which stack do we use, and why does it depend on the project?
We use a usual six-layer stack and each project takes only what it needs: a chatbot with RAG does not need the same as a taxi app with payments and real time.
We call it technology without religion: we decide based on what the product does, what it connects to and who will maintain it.
| Layer | Technologies | When we choose it |
|---|---|---|
| Frontend | Angular, Ionic, TypeScript, Sass; WordPress and Elementor | Dashboards, SaaS and apps with one codebase; WordPress for websites your team edits. |
| Backend | Node.js, NestJS, Python, FastAPI, Firebase, REST APIs, microservices | NestJS for real-time APIs; FastAPI when AI and data weigh most; Firebase to launch fast. |
| Data | PostgreSQL, MongoDB, Redis, Firestore | Depending on the type of data: transactions, documents, cache or synchronisation. |
| AI | OpenAI, Anthropic, Gemini, DeepSeek, Perplexity, RAG, embeddings, agents | By quality, cost, speed, privacy and type of task; RAG to answer from your documents. |
| Integrations | Stripe, PayPal, WhatsApp Business API, Telegram Bot API, Google Maps, Google Sheets, Firebase Auth, Mailgun, Twilio, Telnyx | Payments, messaging, email, maps, authentication, calls and SMS. |
| Infrastructure | VPS, cloud, Google Cloud, Docker, CI/CD, monitoring | From the first deployment, depending on cost, control and operations. |
Frontend and apps
Why Angular and Ionic for web and apps?
Because with Ionic and Angular a single TypeScript codebase publishes the app on iOS and Android, and the same framework serves the website and the admin panel.
Angular gives structure and typing to applications that keep growing. Ionic turns them into apps for the app stores, with access to geolocation, push notifications or Bluetooth.
In 23TopTraining, a training app built with Ionic, Angular and Firebase, the single codebase halves the maintenance cost. Mallorcab has a customer app and a driver app, with a Bluetooth connection to the taximeter and real time. In Scouz, an amateur football app with video, Digitup took part in the development with Ionic and Angular.
If your app needs purely native development, such as a 3D game, we are not the right choice and we will tell you on the first call. More in mobile app development.
Backend
Node.js with NestJS or Python with FastAPI?
It depends on what weighs most: we usually pick Node.js with NestJS for APIs with business logic and real time; Python with FastAPI when the core is AI or data, and Firebase to launch fast without managing servers.
| Criterion | Node.js + NestJS | Python + FastAPI | Firebase |
|---|---|---|---|
| We choose it when | There are permissions, states and real time | AI or documents are at the centre | An MVP or an app has to launch fast |
| Strong point | The same TypeScript as Angular | Python's AI and data ecosystem | Authentication, data and push already managed |
| Watch out for | Keeping heavy data processes separate | A second language next to the frontend's TypeScript | Usage-based cost and provider dependence |
They are not mutually exclusive: the API can run on NestJS and the AI in a Python service. We split into microservices only for a clear reason. Mallorcab uses NestJS; 23TopTraining and Rescrito, Firebase. We apply this in custom software and SaaS development.
Data
PostgreSQL, MongoDB or Firebase?
It depends on the type of data, and in a large product several coexist: Mallorcab combines PostgreSQL, MongoDB and Redis.
| Database | We choose it for | Published project |
|---|---|---|
| PostgreSQL | Relational data and transactions: payments, invoices, permissions | Mallorcab |
| MongoDB | Documents with a variable structure | Mallorcab |
| Redis | Cache and real time | Mallorcab |
| Firestore | Apps and MVPs on Firebase | 23TopTraining, Rescrito |
Firebase speeds up a first version without setting up servers, in exchange for depending on a provider and for a cost that grows with document reads, writes and deletes. More in MVP development for startups.
Artificial intelligence
Which AI models do we use and how do we choose them?
We integrate OpenAI, Anthropic, Gemini, DeepSeek, Perplexity or others depending on quality, cost, speed, privacy and type of task. There is no single model for every project.
We decide with a prototype that validates answers and costs on real examples from your case: a large model can be overkill for classifying short messages, and Perplexity fits when you need web search with citations. Rescrito combines OpenAI, Gemini and other LLMs, and the assistant on this website runs on Gemini on Google Cloud.
The model is only one piece: RAG and embeddings count just as much, to answer from your documents without retraining models, as do agents that query systems and human approval where needed. More in AI agents, AI integration and RAG for business.
Infrastructure
Where is it deployed and how is it operated?
On a VPS or on cloud services such as Google Cloud or Firebase, depending on cost, control and operations, with Docker, CI/CD and monitoring from the first deployment.
A VPS gives a fixed cost and control; managed services reduce operations in exchange for a variable cost. Mallorcab, for example, runs on a VPS with CI/CD and ongoing maintenance.
Operating includes access control, logs, credential protection and backups. If you prefer us to operate it, Build & Run includes hosting, security, support and an evolution allowance. More in software maintenance and ways of working.
Existing product
Can you work with the stack we already have?
We review it on the first call, before committing to anything: if it uses these technologies, we can evolve it; if not, we will tell you frankly whether we are the right team.
We step into platforms that are already live to organise the architecture, improve performance, add AI or prepare for scale. Often the sensible move is to keep the core and integrate the new work through an API: if your ERP works, it stays.
Real stacks
What stack are our projects built on?
Mallorcab and 23TopTraining use Ionic and Angular; Rescrito, Angular and Firebase, and Mundomar, Apache Velocity templates with RAG. All projects are in our case studies.
Mallorcab: taxi platform
- Apps: Ionic, Angular and TypeScript
- Backend: Node.js, NestJS and Firebase Auth
- Data: PostgreSQL, MongoDB and Redis
- AI: OpenAI and Gemini
- Stripe, Google Maps, WebSockets, CI/CD and VPS
Rescrito (formerly Aithor): AI SaaS
- Angular and Firebase (Auth, Firestore and cloud functions)
- OpenAI, Gemini and other LLMs, with RAG
- Stripe, PayPal and Mailgun
23TopTraining: training app
- Ionic and Angular for iOS and Android
- Firebase Auth, Cloud Firestore and push notifications
- Back office for the trainer
Mundomar Cruceros: landing pages with AI
- HTML5, CSS3 and JavaScript
- Gulp and Apache Velocity templates
- LLM and RAG over the cruise catalogue
Fit
When does our stack fit and when does it not?
It fits web, hybrid apps and AI connected to your systems; it does not fit if you need purely native development or a stack that does not appear here.
When it fits
- You want a website, dashboard and apps on a single codebase.
- You need a backend that connects payments, messaging, AI and your systems.
- You want to choose the AI model by results, without being tied to a provider.
- Your product already uses these technologies and you need to evolve it.
When it does not
- Your app needs purely native development: our mobile approach is hybrid.
- You are looking for specialists in a stack that does not appear on this page.
- You want to rent individual developers without a product brief.
- The budget is below €1,000 or the delivery is under 2 weeks.
Related services
Where we apply this stack.
Frequently asked questions about our stack
Which technologies does Digitup Studio use?
Digitup Studio, a technical studio in Madrid, works with Angular, Ionic and TypeScript for frontend and apps; Node.js, NestJS, Python, FastAPI and Firebase for backend; PostgreSQL, MongoDB, Redis and Firestore for data; OpenAI, Anthropic, Gemini, DeepSeek and Perplexity for AI, with RAG, embeddings and agents; and VPS, cloud, Docker and CI/CD for deployment. For content websites it uses WordPress with Elementor.
Why Ionic instead of two native apps?
Because with Ionic and Angular a single TypeScript codebase publishes the app on iOS and Android, and every change is coded once. In 23TopTraining, a training app published on the App Store and Google Play, that single codebase halves the maintenance cost. If your app needs purely native development, such as a 3D game, we are not the right choice and we will tell you on the first call.
Which AI model will you use in my project?
The one that best solves your task. We integrate OpenAI, Anthropic, Gemini, DeepSeek, Perplexity or others depending on quality, cost, speed, privacy and type of task, and we decide with a prototype that validates answers and costs on real examples from your case. We do not tie you to a provider: the integration lets you switch models or combine several if another works better or costs less.
Can you continue a project built by another team?
Yes, if its stack is reasonable to maintain. On the first call, 30 minutes with no commitment, we look at which technologies it uses. Then we review code, infrastructure and deployment, and tell you whether it makes sense to evolve what exists or rebuild some part; if the core works, the usual choice is to keep it and integrate the new work through an API. If it uses a stack we do not work with, we will tell you frankly.
Will the code and infrastructure be ours?
It depends on the way of working you choose. With a fixed-scope project, the software and access are handed over when the project is completed. With flexible payment, the investment is spread over 6 or 12 months and ownership passes when the payments are complete. With Build & Run you pay a monthly fee that includes hosting, security, support and an evolution allowance, with a purchase option defined from the start.
Do you build WordPress websites?
Yes, when the site is content-driven and your team edits it: we use WordPress with Elementor and, if needed, custom plugins. For the Soñar Despierto foundation, Digitup built and maintains a website with volunteering and donation plugins using Stripe, a chatbot as a plugin and Google Sheets integration. If the job is only a WordPress theme, with no product to build, we are not the right choice.
Which stack does your project need?
Tell us what you want to build or what you already have in production. In 30 minutes with Anton, our CTO, we will tell you which technology we would use and why.