Common Pipeline Architectures

Below is a table for the some common architectures for basic pipelines describing what models and nodes to use

Task	LLM Model	Nodes to use
Document Q/A	GPT-3.5-Turbo or Claude-instant	Input, Output, Semantic search, LLM (fill in system and prompt fields)
Chatbot	GPT-3.5-Turbo	Input, Output, LLM (fill in system and prompt fields), Chat Memory
Perform operations on data	GPT 4-Turbo	Input, Output, CSV Query Loader
Search the internet	GPT-3.5-Turbo or Claude-instant	Input, Output, SerpAPI data loader, Semantic Search, LLM (fill in system and prompt fields)
Create research reports on documents	GPT 4-Turbo	Input, Output, File Loader, Semantic Search, LLM (fill in system and prompt fields)