Introducing deepset Studio

As AI use cases continue to advance, the underlying architectures powering these innovations have grown increasingly complex. Designing intricate Compound AI systems that can handle evolving demands using an ever-expanding array of models, components, and tools is challenging in traditional programming environments. 

At deepset, our mission is to simplify the AI development lifecycle and help companies build composable, modular, and flexible AI solutions for real-world applications. Today, we’re proud to introduce our latest innovation, deepset Studio, a visual programming interface that empowers AI engineers to push the boundaries of AI application development while keeping cognitive load at a minimum.

What is deepset Studio?

deepset Studio is built for Haystack, our popular open-source AI framework. Any developer, data scientist, product manager, or aspiring AI engineer can use deepset Studio for free to architect a wide range of LLM applications. Through a user-friendly drag-and-drop UX, users can:

1. Access Haystack’s comprehensive library of components and integrations.
2. See all components and their properties at a glance.
3. Combine them into ever more powerful pipelines. 
4. Visualize their architecture for better reasoning and communication with stakeholders.
5. Efficiently switch between 1:1 code and visual views.
6. Export the final setup as a YAML file (and soon Python code) for use in different environments.

How deepset Studio helps AI teams

deepset Studio helps organizations share AI learnings and progress through visual representations of complex AI systems. It was developed with the AI team in mind – a cross-functional group with varying technical backgrounds. The graphical user interface allows all of those team members – be they from product, engineering, data, or stakeholder teams – to architect, showcase, and modify AI pipelines that power agentic and RAG applications.

The advantages of working in deepset Studio

deepset Studio drives significant developer efficiency, reducing the time to iterate and prototype. Users can draw from Haystack's existing library of integrations and components to create flexible application architecture while relying on deepset Studio’s automatic validation of component relationships and pipeline structure for code quality. They can also jumpstart the development process with proven pipeline templates and component configurations or build their own. An AI engineer can opt to keep writing code in their IDE but when they architect and debug their AI system, they rely on deepset Studio leading to more productivity and developer happiness.

What’s more, deepset Studio comes with multiple serialization options. Any AI pipeline setup designed in deepset Studio can be exported in YAML format (and soon Python) for use in existing development workflows. deepset Studio is natively integrated within deepset Cloud, and users can easily deploy AI pipelines from directly Studio in one click on secure cloud infrastructure.

Optimize deployments with NVIDIA AI Enterprise

As part of our launch, we’re proud to announce a native integration with NVIDIA AI Enterprise software. NVIDIA AI Enterprise users can optimize their deployments through deepset Studio’s integration with NVIDIA NIM microservices and the NVIDIA API catalog. Users can configure NIM microservice deployments and LLM inference directly in Studio. Check out our deployment guide for setting up NIM and Haystack pipelines on Kubernetes, streamlining deployment to any cloud or data center.

Sign up for our waitlist

Any developer can register and use deepset Studio for free, unlocking an interactive visual environment to learn and explore Haystack components and pipelines, and ultimately build LLM applications. The deepset team will continue to enhance deepset Studio to make it the preferred choice for Haystack developers to build pipelines. deepset Studio is currently in beta. To get access, join our waitlist here.