Build a PDF Document RAG System with Mistral OCR, Qdrant and Gemini AI

by Davide Boizza ✓ · 48,030 views · created 2025-05-26 · Document Extraction, AI RAG

What this workflow does

This AI workflow has 29 nodes and runs manually, when you click “Execute workflow” (plus 2 other triggers). Along the way it connects Embeddings OpenAI, Google Gemini Chat Model, Qdrant Vector Store and Google Drive. It also calls api.mistral.ai and qdranturl over HTTP.

How it runs

Trigger: Manually, when you click “Execute workflow” (plus 2 other triggers).

1. Refresh collection (HTTP Request): calls an HTTP API (POST)
2. Get PDF (Google Drive): download
3. Question and Answer Chain: answers questions over your documents (RAG), with model: Google Gemini Chat Model; retriever: Vector Store Retriever
4. Search PDFs (Google Drive): work with file folder
5. Mistral Upload (HTTP Request): calls api.mistral.ai (POST)
6. Loop Over Items (Loop Over Items (Split in Batches)): loops over the items in batches
7. Mistral Signed URL (HTTP Request): calls an HTTP API
8. Edit Fields (Edit Fields (Set)): sets 1 field: file_id
9. Mistral DOC OCR (HTTP Request): calls api.mistral.ai (POST)
10. Execute Workflow (Execute Sub-workflow): calls another n8n workflow

…plus 8 more steps. See the full node list below.

Services it uses

Embeddings OpenAIGoogle Gemini Chat ModelQdrant Vector StoreGoogle Drive

HTTP calls to: api.mistral.ai, qdranturl

Before you upgrade to n8n 3.0

Nodes used (18)

• HTTP Requestn8n-nodes-base.httpRequest
• Manual Triggern8n-nodes-base.manualTrigger
• Loop Over Items (Split in Batches)n8n-nodes-base.splitInBatches
• Embeddings OpenAI@n8n/n8n-nodes-langchain.embeddingsOpenAi
• Default Data Loader@n8n/n8n-nodes-langchain.documentDefaultDataLoader
• Token Splitter@n8n/n8n-nodes-langchain.textSplitterTokenSplitter
• Chat Trigger@n8n/n8n-nodes-langchain.chatTrigger
• Question and Answer Chain@n8n/n8n-nodes-langchain.chainRetrievalQa
• Google Gemini Chat Model@n8n/n8n-nodes-langchain.lmChatGoogleGemini
• Vector Store Retriever@n8n/n8n-nodes-langchain.retrieverVectorStore
• Qdrant Vector Store@n8n/n8n-nodes-langchain.vectorStoreQdrant
• Coden8n-nodes-base.code
• Waitn8n-nodes-base.wait
• Execute Sub-workflown8n-nodes-base.executeWorkflow
• Execute Workflow Triggern8n-nodes-base.executeWorkflowTrigger
• Edit Fields (Set)n8n-nodes-base.set
• Summarization Chain@n8n/n8n-nodes-langchain.chainSummarization
• Google Driven8n-nodes-base.googleDrive

From the author

This workflow is designed to process PDF documents using Mistral's OCR capabilities, store the extracted text in a Qdrant vector database, and enable Retrieval-Augmented Generation (RAG) for answering questions. Here’s how it functions: Once configured, the…

View the original on n8n.io → Check your own copy