dashAI

dashAI is a local-first, open-source desktop application for data analysis and machine learning model training. It allows users to import datasets, train predictive and generative models, evaluate performance, and apply explainability techniques entirely on their own machines without requiring cloud services, external authentication, or API keys.

dashAI is suitable for researchers, data scientists, and developers who prefer working with local data and models. Its schema-driven interface and modular architecture make it accessible to users who want to avoid cloud dependencies while maintaining flexibility and extensibility.

dashAI provides an integrated pipeline for data exploration, model training, evaluation, and interpretability through a single visual interface. Models are registered as subclasses of twelve base classes, ensuring a consistent user experience across tasks such as tabular classification, NLP, translation, and text-to-image generation.

No, dashAI allows users to train machine learning models without writing code. The interface is schema-driven and automatically generated from Pydantic schemas, eliminating the need for manual frontend development or coding.

Yes, dashAI supports extensibility through a modular architecture. Users can install plugins directly from the UI or PyPI without restarting the application or using external shells. The project encourages community contributions and provides a plugin marketplace.

Users can start by downloading the desktop application from the official website. Once installed, they can import datasets, explore models in the built-in catalog, and begin training or evaluating models through the visual interface. Documentation and community resources are available for guidance.