|Index|Using SUSE AI Factory|What pre-built AI blueprints are included?
Applies to SUSE AI Factory

3 What pre-built AI blueprints are included?

SUSE AI Factory includes the following pre-built AI blueprints:

3.1 Simple chatbot with RAG

This blueprint features a simple chatbot with retrieval-augmented generation (RAG) using Ollama, Open WebUI and Open WebUI-mcpo. It uses CPU inference and is suitable for proofs of concept (PoCs) and quick explorations. RAG uses the ChromaDB vector store embedded in Open WebUI.

3.2 Simple chatbot with RAG (vLLM and Milvus)

This blueprint features a simple chatbot with RAG using vLLM, Open WebUI, Milvus and Open WebUI-mcpo. It is suitable for proofs of concept (PoCs) and beyond. vLLM serves the large language model (LLM) over an OpenAI-compatible API and requires NVIDIA GPU nodes. The serving engine requests one GPU and uses the NVIDIA runtime class.

3.3 Simple inference endpoint

This blueprint is a turnkey production inference LLM gateway that fronts one or more vLLM model back-ends behind a single OpenAI-compatible API. It adds authentication, role-based access control (RBAC) and optional prompt-injection guardrails. All components run on SUSE images.