GenAI beyond basics

A collection of Generate AI related samples.

Samples

Frameworks

LangChain

• Chat with in-memory history
• Chat with history saved to Firestore

Semantic Kernel

• Chat with in-memory history

Grounding & RAG

Grounding

• Grounding with public data with Google Search
• Grounding with your own data with Vertex AI Search

RAG

• RAG with a PDF using LangChain and Annoy Vector Store
• RAG with a PDF using LangChain and Firestore Vector Store
• RAG with a PDF using LlamaIndex and SimpleVectorStore on Vertex AI
• LlamaIndex on Vertex AI with RAG API
• RAG with a PDF using File Search Tool in Gemini API

Retrieval

• Multimodal image storage and retrieval with Chroma
• Multimodal image storage and retrieval with Firestore Vector Store

Function calling

• Function calling - Weather

Input and output

• Response type and schema (Vertex AI)
• Response schema and pydantic (LangChain)
• Context caching
• Batch generation of content

Evaluation

• Gen AI evaluation service in Vertex AI
• DeepEval and Gemini
• Promptfoo and Vertex AI

Guardrails

• Model Armor
• LLM Guard and Vertex AI

Tracing

• Tracing with Langtrace and Gemini

Protocols

• Model Context Protocol (MCP)
• Agent2Agent Protocol (A2A)
• Agent-User Interaction Protocol (AG-UI)
• Agent to UI Protocol (A2UI)