Web and frontend
Corporate sites, portals and SaaS that load fast, rank well and stay easy to maintain.
- React
- JavaScript
- TypeScript
- Vite
- HTML5/CSS3
Claude through the Anthropic API, Azure, SQL Server and C#. It is what we run in production, in our own businesses and in our clients' businesses.
There are four pieces. All four work together in Recepción24.
Within Claude, we pick the model that makes the fewest mistakes where a mistake costs the customer something, even when it is the more expensive one. The agent does more than talk: it uses tools that connect it to the company's systems, to look up an order, check a calendar or open a ticket. The answer is generated as a stream, which is what lets an agent reply on the phone at the pace of a conversation. And the fixed instructions of each agent are cached, so every conversation costs less.
The applications run on Azure App Service, in European data centres. Every change goes live automatically with GitHub Actions. Azure Storage queues hold pending work so that none gets lost, and Application Insights logs the errors. Cloudflare filters the traffic before it reaches the applications.
A relational database keeps the conversations, what each one cost, and the business data. It is proven technology, with automatic backups and with queries any engineer can read.
The programs that drive the agents, the APIs and the control panels are written in C#, on .NET 10. The screens are built with Blazor. In the voice agents we also use Python, with LiveKit.
Corporate sites, portals and SaaS that load fast, rank well and stay easy to maintain.
Native when performance rules, cross-platform when the deadline does. We know what breaks when an app grows.
Node.js, PHP or PostgreSQL when the project is already built with them. AWS, Docker and nginx when the client already has them.
Your ERP and your CRM talking to each other, with no copy and paste. If it has an API, we connect it.
When a CMS solves the problem, we use it and save you months. We do not rewrite for sport.
Measure what works and show up where people look for you: on Google and in the AI assistants.
The everyday stack is the one above. If yours is a different one, these four rules decide:
AI receptionist
Recepción24
Engine in C#, with Claude and Azure SQL, on Azure App Service.
See case study →
Pricing for short-term rentals
ListingOK
Application in Blazor and .NET, with Azure SQL, on Azure App Service.
See case study →
Home repairs
Reparaman
Website in Blazor and .NET 10, on Azure App Service. Its receptionist is a Recepción24 AI agent.
See case study →If you want to talk to us, the first call is free. Maybe something comes of it, maybe not, but you will know more by the time we hang up.
Let’s talk →