Getting started with FABRIC just became easier.
We're excited to announce the availability of LoomAI, a new hosted, AI-augmented graphical interface designed to simplify how researchers create and manage experiments on the FABRIC testbed. By combining an intuitive visual editor with conversational AI assistance, LoomAI lowers the barrier to entry for new users while helping experienced researchers build experiments more efficiently.
Unlike traditional workflows that require familiarity with the FABRIC Python SDK, LoomAI enables users to describe the experiment they want to create in natural language. The integrated AI assistant translates those requests into experiment configurations, making it easier to design, modify, and deploy slices without extensive coding experience.
LoomAI complements, not replaces, the existing FABRIC Portal and JupyterHub environment. Users can continue working with their existing notebooks while taking advantage of the new graphical interface. JupyterHub remains fully integrated, allowing researchers to seamlessly access and continue their current workflows.
The hosted version of LoomAI is now available to FABRIC users with valid project credentials at https://loomai.fabric-testbed.net/hub/login and on the FABRIC Portal. Access permissions continue to follow each user's existing project memberships and resource allocations, ensuring a familiar and secure experience.
In addition to supporting conversational experiment creation, LoomAI provides access to multiple AI models. Users can leverage FABRIC-hosted open-weight models, including Gemma and Qwen3, with provided credentials. Support for external AI services such as OpenAI, Claude, and Codex is also available for users who wish to connect using their own API credentials.
Looking ahead, LoomAI is being developed as a unified interface that extends beyond FABRIC. Integration efforts are already underway with additional research testbeds, including support for cross-testbed experimentation with platforms such as Chameleon and Sphere. The long-term goal is to enable researchers to design experiments spanning multiple cyberinfrastructure platforms from a single workspace.
The hosted platform has already been evaluated by FABRIC REU students and early community users over the past several months, helping refine the interface and user experience before broader availability. As LoomAI continues to evolve, community feedback will play an important role in guiding future enhancements and new features. You can provide feedback by visiting the Help page to access the Knowledge Base, schedule Office Hours, or post on the Forum.
Whether you're new to FABRIC or looking for a faster, more intuitive way to build experiments, LoomAI offers a powerful new way to accelerate research. To learn more about LoomAI and see it in action, explore the recorded webinar and stay tuned for additional documentation and resources as development continues.