Build an AI agent that answers user questions from documents
Workflow Description
Advanced workflow that integrates a language model with a vector database to process natural language queries. Loads documents, splits them, creates embeddings, stores them in Qdrant, then routes questions to an agent that retrieves relevant answers with context.
How it works
- 1.Load documents and split them into processable text chunks using recursive character splitting
- 2.Convert text segments into numerical vectors using Mistral embeddings
- 3.Store vectors in Qdrant vector database for efficient semantic search
- 4.Receive user inquiries through chat interface and pass them to the intelligent agent
- 5.Extract relevant context from vectors and deliver accurate answers to the user
Use cases
- Customer support system that answers questions from internal knowledge bases
- Intelligent search engine that understands user intent and provides accurate responses
- Specialized virtual assistant handling multi-domain queries efficiently
Requirements
- Valid API key for Mistral Cloud embeddings and language model
- Configured Qdrant vector database instance with active connection
- Text documents prepared for upload (PDF, Word, plain text formats)
- Chat Trigger configured for user communication interface
Service Value
Ready-made workflow template for automation delivery and service execution.
Apps Used
Details
How to Use
- 1.Click "Download Template"
- 2.Open your n8n dashboard
- 3.Go to Workflows > Import from File
- 4.Select downloaded file and configure credentials
Nodes Used (33)
When clicking "Test workflow"
Manual Trigger
Get This Week's Menu
HTTP Request
Extract Available Courses
Code
Extract Server Data
HTML Page
Get Course Metadata
Set
Get Recipe
HTTP Request
Embeddings Mistral Cloud
Embeddings Mistral Cloud
Default Data Loader
Document Default Data Loader
Merge Course & Recipe
Merge
Prepare Documents
Set
Recursive Character Text Splitter
Text Splitter Recursive Character Text Splitter
Chat Trigger
Chat Trigger
Extract Recipe Details
HTML Page
Qdrant Recommend API
Tool Workflow
Execute Workflow Trigger
Execute Workflow Trigger
Mistral Cloud Chat Model
Lm Chat Mistral Cloud
Get Tool Response
Set
Wait for Rate Limits
Wait
Get Mistral Embeddings
HTTP Request
Use Qdrant Recommend API
HTTP Request
Get Recipes From DB
Code
Save Recipes to DB
Code
Sticky Note
Sticky Note
Sticky Note1
Sticky Note
Sticky Note2
Sticky Note
Sticky Note3
Sticky Note
Sticky Note4
Sticky Note
AI Agent
Agent
Sticky Note5
Sticky Note
Sticky Note6
Sticky Note
Qdrant Vector Store
Vector Store Qdrant
Sticky Note7
Sticky Note
Sticky Note8
Sticky Note