Retrieval-augmented generation pipelines are the backbone of any LLM-over-your-data app. These are the cleanest ingestion + query patterns in the catalog: chunkers, embedders, retrievers, and the agent loops that stitch them together.
Top 20 RAG Pipelines — ranked by quality
Ranked across 1,069 rag pipelines in the catalog by Pro QualityScore (multi-signal blend: graph topology, error discipline, naming, documentation density, metadata depth). The full ranking is a Pro feature; the top 20 are visible here for free.
⚡AI-Powered YouTube Playlist & Video Summarization and Analysis v2. Uses lmChatGoogleGemini, agent, splitOut, chainLlm. Chat trigger; 72 nodes.
This n8n workflow transforms entire YouTube playlists or single videos into interactive knowledge bases you can chat with. Ask questions and get summaries without needing to watch hours of content. 🔗
AI-Powered YouTube Playlist & Video Summarization and Analysis v2. Uses lmChatGoogleGemini, agent, chainLlm, vectorStoreQdrant. Chat trigger; 72 nodes.
Who is this for? This workflow is ideal for HR teams, startups, and enterprises that want to handle employee interactions through WhatsApp and automate responses using LLM (OpenAI) and intelligent rou
Transform raw investment memorandums and financial decks into comprehensive, professional Due Diligence (DD) PDF reports. This workflow automates document parsing via LlamaParse, enriches internal dat
🤖 AI Powered RAG Chatbot for Your Docs + Google Drive + Gemini + Qdrant. Uses documentDefaultDataLoader, textSplitterTokenSplitter, vectorStoreQdrant, splitInBatches. Event-driven trigger; 50 nodes.
This workflow creates a powerful RAG (Retrieval-Augmented Generation) chatbot that can process, store, and interact with documents from Google Drive using Qdrant vector storage and Google's Gemini AI.
Streamline M&A due diligence with AI. This n8n workflow automatically parses financial documents using LlamaIndex, embeds data into Pinecone, and generates comprehensive, AI-driven reports with GPT-5-
This n8n workflow automates the process of ingesting documents from multiple sources (Google Drive and web forms) into a Qdrant vector database for semantic search capabilities. It handles batch proce
Security incidents. Uses googleDriveTrigger, googleDrive, vectorStorePinecone, documentDefaultDataLoader. Event-driven trigger; 29 nodes.
Dental Clinic – AI Booking Agent (Webhook + HTTP Request / Cloud Station). Uses agent, lmChatGoogleGemini, memoryBufferWindow, googleCalendarTool. Webhook trigger; 27 nodes.
This workflow contains community nodes that are only compatible with the self-hosted version of n8n.
Automatically draft email replies using AI. This workflow monitors your Gmail inbox, filters out automated emails (newsletters, receipts, notifications), and uses AI to create draft responses only for
Vector DB Loader from Google Drive. Uses documentDefaultDataLoader, textSplitterRecursiveCharacterTextSplitter, embeddingsOpenAi, vectorStorePGVector. Event-driven trigger; 15 nodes.
Automatically convert documents from Google Drive into vector embeddings using OpenAI, LangChain, and PGVector — fully automated through n8n.
This workflow ingests incident and playbook content from GitHub into Supabase (including pgvector embeddings with Google Gemini) and, on a webhook trigger, enriches a test incident with historical mat
___MAIN_2_Modularizado_24_3___ATIVO___ADVOCACIA_n8n___AGENTE_ENCAMINHAMENTO_. Uses supabase, httpRequest, informationExtractor, lmChatOpenAi. Event-driven trigger; 68 nodes.
Chat with docs - 5minAI New version. Uses httpRequest, documentDefaultDataLoader, textSplitterRecursiveCharacterTextSplitter, embeddingsOpenAi. Event-driven trigger; 62 nodes.
project Dick n8n. Uses agent, memoryBufferWindow, mistralAi, lmChatOllama. Webhook trigger; 60 nodes.
Iroko-chatbot-green-API. Uses toolCalculator, agent, n8n-nodes-whatsapp-green-api, chatTrigger. Webhook trigger; 54 nodes.
What makes a workflow rank in this list
The QualityScore is calibrated against a hand-rated ground truth of 20 workflows (Spearman ρ = +0.941). Workflows score higher when they have: clean graph topology (no spaghetti), explicit error branches, sticky-note documentation, complete parameter coverage, and rich metadata (description, integrations, credentials). See the methodology page for the full axis breakdown.