DashAI

DashAI is an open source desktop application for data analysis and machine learning model training without writing code. It runs entirely on the user's own machine and needs no cloud services, external authentication or API keys for its core functions. Within one visual interface users import datasets, explore data, train predictive and generative models, evaluate them and apply explainability techniques. It suits students, researchers and analysts who want to experiment with machine learning on local data in an environment they control. The catalog lists 76 models: 30 tabular models, 15 natural language processing models and 9 translation models on the predictive side, plus 5 open weights large language models and 11 text to image models, among them Llama, Mistral, Qwen, PixArt Sigma and SDXL. The architecture rests on twelve typed base classes for registering models, data loaders, metrics, optimizers and explainers. The interface is generated from each component's Pydantic schema, so a new component appears in the UI without any frontend code, and community plugins install from PyPI with one click inside the application. DashAI is free. The code is open source with no copyleft clauses and no reserved proprietary components, and it can be downloaded from the website, GitHub, PyPI or SourceForge. DashAI is a Chilean academic project maintained by a network of research centers with public funding rather than a single company. The current release is still labelled alpha, and the website is primarily in Spanish.