Welcome to AI4EOSC documentation!¶

The AI4EOSC platform empowers scientists by lowering the barrier to adopt the latest AI models and tools. The platform provides a user-friendly workbench to easily train, deploy, share and monitor AI models. It covers the full ML lifecycle: from model creation and training to deployment and monitoring in production, adhering strictly to FAIR principles for science (fully open-source and portable with no vendor lock-in). Furthermore, AI4EOSC is actively expanding to support Generative AI (GenAI) and agentic workflows for scientific research.

The AI4EOSC Ecosystem¶

The platform is developed and expanded through a collaborative ecosystem of projects, gateways, and scientific communities:

Projects (DEEP-HDC, AI4EOSC, AI4Life, iMagine, EOSC-ARENA, FLUID-AI, GenAI4Earth): EU-funded initiatives that develop and extend the AI4EOSC platform with new capabilities—from core MLOps, distributed training, and composite AI to Generative AI, agentic environments, and FAIR data/model interoperability.
Gateways (AI4EOSC, iMagine, AI4Life, KMD4EOSC, Tutorials): Dedicated instances of the AI4EOSC platform built on top of the shared platform foundation (common API, authentication, and compute/storage infrastructure). Each gateway is customized with specific branding, tools, and domain-specific model catalogs to access dedicated resources for a given community.
Communities & Scientific Domains: Scientific communities group researchers within specific domains (e.g., climate forecasting, genomics and life sciences, aquatic imaging and biodiversity, and interdisciplinary science). One gateway serves a given community, while a single community can leverage several gateways to access the AI models and computing tools relevant to their research.