Getting Started with AI-on-Demand
AI on-Demand (AIoD) is a collaborative platform for AI research. It aims to bring together the entire AI community to access and share reliable, up-to-date AI resources, created in Europe and beyond. By promoting open and reusable AI, AIoD helps accelerate science, power innovation, and ensure responsible use of AI for the benefit of society. AIoD stands out by bringing together a rich and diverse collection of resources, including: - AI Datasets, Models, Experiments, Tools, and Services - Publications, Educational resources, and Research projects - News, Events, Labs, and People Whether you're looking to build an AI application, create a national policy, prepare a lesson, or just learn something new, AIoD is your launchpad. | 8 videos
0 / 8 completed
For AI researchers, teachers and students
Need reliable AI resources for study, research or innovation? You’re in the right place.
This video takes you on a journey through the platform’s key features, designed to unite AI researchers, students, SMEs, tech providers, EU-funded projects and Digital Innovation Hubs. Explore how AIoD fosters collaboration and drives progress in AI research, experimentation, and development. Key highlights:
Pillar 1: Metadata Catalogue – Access over 500,000 AI-related assets in one place.
Pillar 2: Common infrastructure – Share and use computational resources for AI experiments and services.
Pillar 3: Single sign-on – Easy, secure access to all platform tools and content.
Search AI resources
- Search AI resources through any topic or keyword (e.g. “climate”, “robotics”, “translation”).
- Filter by type (dataset, model, paper, …).
- Access resources available on top platforms like Zenodo, Hugging Face, and OpenML.
- Coming soon: an AI chatbot to help you find resources and explain how to use them.
Explore resources
- Each result links to a detailed summary: what it is, where it comes from, who made it.
- Includes tags, categories, and direct links to download or view on the original site.
- Quickly see how a dataset, model, and publication connect.
Stay organised
- Log in to bookmark your favourite resources (under the AI Catalogue).
Try AI services
AIoD also gives you access to interactive AI services:
– RAIL – Run experiments with AI models, evaluate them, and compare results, all in a reproducible and shareable way.
– AI-Builder – A no-code tool that lets you build machine learning models step-by-step, perfect for non-programmers or educators.
0 / 8 completed
FAQs
The AIoD platform is a collaborative, community-driven digital space that supports European research and innovation in Artificial Intelligence (AI), while promoting the European values of quality, trustworthiness, and explainability.
The AIoD platform is a collaborative, community-driven digital space that supports European research and innovation in Artificial Intelligence (AI), while promoting the European values of quality, trustworthiness, and explainability.
The platform is open to everyone interested in AI, including researchers, SMEs, tech providers, students, EU-funded projects, Digital Innovation Hubs, and public institutions.
The platform is open to everyone interested in AI, including researchers, SMEs, tech providers, students, EU-funded projects, Digital Innovation Hubs, and public institutions.
By joining the AIoD platform, users can access AI-related tools, services, and datasets, share research outcomes, updates, and events, take part in educational courses and training materials, connect and collaborate with other experts and peers and contribute to the development and evolution of the platform.
By joining the AIoD platform, users can access AI-related tools, services, and datasets, share research outcomes, updates, and events, take part in educational courses and training materials, connect and collaborate with other experts and peers and contribute to the development and evolution of the platform.
There are many ways to contribute: from developing technical components or sharing research outputs, to participating in working groups.
To get involved, send an email to info@aiod.eu and a representative from the AI4Europe project will guide you to the right initiative.
There are many ways to contribute: from developing technical components or sharing research outputs, to participating in working groups.
To get involved, send an email to info@aiod.eu and a representative from the AI4Europe project will guide you to the right initiative.
Not at all. While it has a strong research foundation, AIoD also serves industry professionals, startups, students, and public organizations interested in leveraging or contributing to AI.
Not at all. While it has a strong research foundation, AIoD also serves industry professionals, startups, students, and public organizations interested in leveraging or contributing to AI.
You can find datasets, AI tools and services, use cases, scientific publications, funding opportunities, training resources, and upcoming AI-related events.
You can find datasets, AI tools and services, use cases, scientific publications, funding opportunities, training resources, and upcoming AI-related events.
AIoD directly contributes to the European AI strategy by enabling open and trustworthy AI research and development, promoting collaboration and transparency, and ensuring alignment with European ethical and technical standards.
AIoD directly contributes to the European AI strategy by enabling open and trustworthy AI research and development, promoting collaboration and transparency, and ensuring alignment with European ethical and technical standards.
Yes. AIoD also serves as a dissemination channel where users can share research results, news, and announcements about upcoming events by filling in AIoD’s Communication requests.
Yes. AIoD also serves as a dissemination channel where users can share research results, news, and announcements about upcoming events by filling in AIoD’s Communication requests.
Not as such. The example at https://aiondemand.github.io/aiondemand/examples/sharing/ shows how you can upload assets using Python. You could use the REST API interface in a similar fashion if you are not working with Python. To use this process in CI/CD, you would need to log in once manually to obtain a token, then provide this token to CI/CD.
Not as such. The example at https://aiondemand.github.io/aiondemand/examples/sharing/ shows how you can upload assets using Python. You could use the REST API interface in a similar fashion if you are not working with Python. To use this process in CI/CD, you would need to log in once manually to obtain a token, then provide this token to CI/CD.
No, ECAS is not among the federated identity providers.
No, ECAS is not among the federated identity providers.
Yes, you can. If an institution is not yet available, it can be integrated into the system through federation.
Yes, you can. If an institution is not yet available, it can be integrated into the system through federation.
Yes. We are currently storing metadata in AI-on-Demand and actively expanding indexed resources. We currently index datasets from Zenodo, Hugging Face, and OpenML. We plan to add other Zenodo artifact types and a GitHub connection soon. If a connector for a specific asset type or platform is missing, please inform us. In the meantime, you can use our editor tool to manually enter the metadata and location of your asset. As an open-source project, we welcome contributions to this effort.
Yes. We are currently storing metadata in AI-on-Demand and actively expanding indexed resources. We currently index datasets from Zenodo, Hugging Face, and OpenML. We plan to add other Zenodo artifact types and a GitHub connection soon. If a connector for a specific asset type or platform is missing, please inform us. In the meantime, you can use our editor tool to manually enter the metadata and location of your asset. As an open-source project, we welcome contributions to this effort.
No, we welcome all AI-related contributions on the platform.
No, we welcome all AI-related contributions on the platform.
AI-on-Demand provides a web portal with a friendly user interface that does not require any coding at all.
AI-on-Demand provides a web portal with a friendly user interface that does not require any coding at all.
For metadata uploaded directly to the platform, we employ a review process. For metadata indexed from other platforms, we rely on their curation process. It is always possible to report malicious assets for removal.
For metadata uploaded directly to the platform, we employ a review process. For metadata indexed from other platforms, we rely on their curation process. It is always possible to report malicious assets for removal.
Try logging in. If you have it, it will be detected automatically.
Try logging in. If you have it, it will be detected automatically.
You can link multiple identities to a single account as long as you still have access to them.
You can link multiple identities to a single account as long as you still have access to them.
Yes, the connectors will periodically revisit indexed assets to update the metadata on AI-on-Demand.
Yes, the connectors will periodically revisit indexed assets to update the metadata on AI-on-Demand.
Yes, the platform is entirely open source. Find us at https://github.com/aiondemand. Our roadmap is publicly available (Projects > Roadmap) and the various repositories include our server back end, connectors to different platforms, Python connector, and even our metadata schema. We welcome contributions, feedback, bug reports, etc.
Yes, the platform is entirely open source. Find us at https://github.com/aiondemand. Our roadmap is publicly available (Projects > Roadmap) and the various repositories include our server back end, connectors to different platforms, Python connector, and even our metadata schema. We welcome contributions, feedback, bug reports, etc.
Yes. The AI catalogue has a “bookmark” icon (the blue circle with a bookmark outline) that you can click to save a resource for later. It will be accessible from the “My Bookmarks” menu in the AI catalogue. This functionality is also available through the API, so when building your own script or service you will be able to allow for bookmarking or showing bookmarked assets too!
Yes. The AI catalogue has a “bookmark” icon (the blue circle with a bookmark outline) that you can click to save a resource for later. It will be accessible from the “My Bookmarks” menu in the AI catalogue. This functionality is also available through the API, so when building your own script or service you will be able to allow for bookmarking or showing bookmarked assets too!

