Pioneers in Computational Linguistics since 2014
We use artificial intelligence to help you solve real business problems.
For over a decade, we have studied how machines understand human language. From developing proprietary semantic search engines to creating advanced RAG systems, we turn your unstructured data into actionable knowledge.
Excellence built over time
Over 10 years of innovation in the field of AI.
The origins
Launch of Ludwig.guru, our first adaptive semantic search engine. This was our first project on how machines "understand" people.
Neural era
We trained our first proprietary neural paraphraser model, successfully integrating it into the Ludwig ecosystem.
European research
We won an EU-funded grant on Prompt Tuning, specializing in techniques to optimize LLM training at reduced costs.
RAG and generative AI
Evolution toward RAG, combining our expertise in semantic search with the generative power of AI.
RAG according to Futura
Our strongest specialization is Information Retrieval (IR), and we have Ludwig's search algorithm — refined over 10 years of testing on millions of queries — as the engine powering Artificial Intelligence.
Hallucination minimization
The system responds based exclusively on the provided company documents, eliminating fabricated answers.
Accurate data
Thanks to Library (our proprietary system), we enable indexing of complex databases for instant searches.
Privacy and control
Your data doesn't "train" public models; it stays within your security perimeter.
edit.legal: AI at the service of law.
edit.legal is the AI platform designed exclusively for Italian law. Access to over 500,000 court decisions and 1 million regularly updated legal sources, with natural language search and citation precision.
NCP Companion: AI for European funding.
An intelligent virtual assistant developed for FECYT (Spanish Foundation for Science and Technology). It uses our proprietary RAG technology to provide accurate and verifiable answers on European Commission documentation and Horizon Europe guidelines.
PinkWhale: it all starts here.
As the saying goes: "garbage in, garbage out." If the input data isn't extracted correctly, the downstream results will be compromised. Through the PinkWhale project, we developed proprietary solutions for extracting structured data (i.e., computer-usable) from unstructured data (i.e., PDFs, receipts, invoices, bills of lading, etc.).
