The University of Chile unveiled dashAI, a Chilean platform that allows users to work without sharing data

The Faculty of Physical and Mathematical Sciences at the University of Chile, through its Data and Artificial Intelligence Initiative (IDIA), has launched free and open-source software that allows users to train artificial intelligence models without needing to know how to program. The tool operates 100% locally on the user’s computer, without any mandatory reliance on the cloud, positioning the institution at the forefront of public technological development in the country.

The open-source dashAI platform is designed to enable the training and evaluation of machine learning models and the use of generative AI models without the need for programming, and it runs on each user's computer.

The initiative addresses a key challenge in today’s technology landscape: access to artificial intelligence and data science. Creating or evaluating AI algorithms used to require advanced programming skills in Python or reliance on expensive commercial cloud infrastructure; dashAI breaks this paradigm with an intuitive “no-code” interface, and all processing takes place on the user’s computer, so the data never leaves their device.

“The dashAI platform is proof that Chilean academia can develop artificial intelligence technology in an open and collaborative manner, without relying on a single company, private platforms, or foreign infrastructure. For our School of Physical and Mathematical Sciences (FCFM), this represents a demonstration of leadership in AI development to address the country’s various challenges, showing that the democratization of technology requires open tools that do not exclude those who do not know how to code,” said Dean Susana Mondschein. 

The development of this platform was led by the FCFM through its Department of Computer Science (DCC) and the Data and Artificial Intelligence Initiative (IDIA). The creation of this ecosystem involves the National Center for Artificial Intelligence (CENIA) as the beneficiary institution,the Institute for Mathematics, Physics, and Data ( Millennium Institute , IMFD)Foundational Research on Data as a collaborator, and the company Unholster as an industry partner. The tool was scaled up thanks to government funding from the National Agency for Research and Development (ANID) through the Fondef IDeA fund.

“This initiative demonstrates that Chile can develop world-class artificial intelligence through collaboration among universities, research centers, and industry. Our goal is to democratize access to these technologies through open-source tools that allow researchers, companies, and institutions to innovate without sacrificing control over and privacy of their data,” explained Felipe Bravo, director of IDIA FCFM, professor at the DCC, IMFD researcher, and project leader .

“Thisplatform has also allowed us to strengthen a technology development model that brings together the efforts of university students with a long-term vision to build technologies as a public good for Chile and the region. More than 20 students working on their theses and master’s theses at the University of Chilehave participated in the process, ” added Claudia López, who is also a faculty member at the FCFM at the University of Chile.

In terms of technical innovation, dashAI functions as a desktop application that unifies data processing, training, evaluation, and explainability for 76 available models —ranging from tabular classifiers to advanced generative models such as Mistral and DeepSeek— into a single environment .

According to Juan Reutter, director of IMFD, this platform“allows us to train AI models in a local environment, maintaining the sovereignty and privacy of trainingdata, and offering a valuable alternative to cloud services. The fact that People can analyze their data and train models while retaining control over their informationis an important step toward fostering more accessible and responsible AI, since without well-managed data, we cannot create reliable models.”

Each tool added to dashAI comes with a description of its own options, and the program automatically generates the screen for configuring it. This makes it possible to incorporate new components with very little additional work. To date, the platform has been downloaded more than 18,500 times worldwide.