Runnable reference architectures for retrieval-augmented generation, streaming-to-search, and agentic infrastructure workflows.
Each lab ships with a sample dataset, discussions on decision points, and the exact steps to run it on your laptop. Get hands on with important use cases and architectures by running it, not by reading about it.
| Lab | You'll build | Repo | Notebook | Videos |
|---|---|---|---|---|
| Retrieval-Augmented Generation | A retrieval application using hybrid search and vector retrieval - load sample data, validate retrieval behavior, understand the pattern. | Getting Started | Coming soon | InstAcademy |
| Agentic RAG | An Agentic RAG agent that implements a Mixture of Experts (MoE) architecture with 2 Experts, OpenSearch as a vector store, and MCP tooling. | Getting Started | Coming soon | Video |
| Streaming-to-Search | An event pipeline that streams operational data into a searchable index - update search state, validate indexing, inspect failure handling. | Coming soon | Coming soon | - |
Each lab also comes with a concise use-case page, architectural walkthrough, and a path to take it further when you're ready to operationalize it. That path shows you what changes past the demo: governed responses, repeatable results, and observability for production readiness. Least-privileged access, audit trails, and approval gates for agentics. That's the stuff that turns a working prototype into something you can actually trust.
Once a lab is live, there will be three ways to work through it, depending on what you're looking for:
- Just want to look? Download the Jupyter notebook and hit "play" on the interactive code.
- Want to build? Run the self-guided, hands-on path on your laptop using Instaclustr's free trial - built to work with AI CLIs like Claude, Cursor, or Codex.
- Want to learn the concepts first? Take an InstaAcademy course (launching September 21, 2026) - videos plus a hands-on lab.
All contributors are welcome - community members, customers, and NetApp employees alike - through pull requests on the public GitHub workflow. Contributions are subject to a CLA and are reviewed and merged at the discretion of the Maintainers listed in MAINTAINERS.md.
New labs get added as the community asks for them and helps validate them.