Visual and Interactive AI Workflows (OrangeWeb)
Overview
OrangeWeb is a user-friendly AI workflow platform based on visual programming. It enables researchers, educators, and practitioners to easily connect data sources, apply advanced AI models, explore results, and understand model behavior without programming expertise. The platform supports rapid prototyping, reproducible experimentation, and explainable AI use across diverse application domains. Furthermore, OrangeWeb's integration with HPC resources allows the use and adaptation of pre-trained deep models. OrangeWeb is built around a visual programming framework in which users compose custom workflows by connecting reusable components, referred to as widgets. These widgets provide capabilities for data access, visualization, exploratory analysis, and modeling across multiple data modalities. The platform is accessed through a web browser, while computational tasks are executed by a Python-based backend deployed on HPC infrastructure. The backend leverages modern AI and machine-learning libraries, including scikit-learn and PyTorch, which are exposed to users through widgets offering interactive visual controls.
Outcome
The service democratizes the development of AI-driven solutions by reducing technical barriers to experimentation and deployment. Users can rapidly prototype and customize AI workflows, reuse standardized components, and gain insights into model behavior. Supported by HPC infrastructure, the service enables efficient adaptation and application of pre-trained deep models to new datasets.
Service Card
| Category | AI development and deployment |
| Domain | Visual Programming, AI Workflow Platform |
| Vertical | Cross-domain |
| Target users | SMEs, Startups, Industrial R&D |
| Service type | Pipeline |
| Outputs | Workflows, models, explanations, visualizations. |
| Requirements | Requires an NVIDIA GPU node for the computation of the back-end and a database node for workflow persistence. |
| Limitations | |
| Availability | Q3/2027; prototype deployed Q3/2026 |
| Licensing | GPL3 |
| Visual content | |
| Visual caption |