Own product · Applied AI
urbanisti.co: planning regulations with verifiable answers
An AI product over Spanish urban-planning regulations for architects and municipal technical staff. I am its founder and work on it from Granada: a concrete RAG application that makes the source part of the answer.
The problem
Consulting planning regulations means finding the relevant document and checking what it says. A convincing answer is not enough: the person using it needs to return to the source and verify its contents. urbanisti.co addresses that document lookup for architects and municipal technical staff.
The solution: RAG over regulations
RAG combines document retrieval with answer generation. urbanisti.co applies it to Spanish planning regulations: the system answers using documentation and includes citations to its sources.
Citations let the reader check the answer against the original text. This supports information lookup; professional interpretation and checking the applicable regulations remain necessary.
My role and technical approach
As its founder, I maintain this RAG system in production. The project brings together backend engineering, information retrieval and language-model integration.
The project uses FastAPI, LangGraph, BigQuery Vector Search and Next.js. FastAPI provides the backend, LangGraph organises the AI workflow, BigQuery Vector Search supports vector retrieval, and Next.js serves the interface.
What you can check
The product is available at urbanisti.co. Visit its website to explore the offering and how it presents regulation lookup.
Applying this to other projects
The same need for answers with sources appears in company manuals, procedures and knowledge bases. Each implementation requires decisions about what information to retrieve, who can access it and how to evaluate answers.
If your business needs this kind of solution, we can start by reviewing your documents and defining a concrete scope. I am based in Granada and work remotely in Spanish and English.