AI-Powered Local Document Q&A Agent with Mistral & Qdrant

Unlock Instant Insights from Your Local Documents with this AI Agent

This AI Agent automates the process of creating and maintaining a searchable knowledge base from files stored in a local directory. It watches for new, updated, or deleted files, processes their content, generates vector embeddings using Mistral AI, and synchronizes them with a Qdrant vector database. Once set up, you can chat with your documents, asking complex questions and receiving AI-generated answers based on the indexed information. This is a powerful tool for solopreneurs, founders, and technical teams looking to quickly leverage proprietary information or build internal Q&A systems.

Key Features & Benefits

• Automated Document Ingestion & Sync: Monitors a local folder and automatically processes file additions, updates, and deletions.
• AI-Powered Indexing: Uses Mistral AI to create dense vector embeddings for efficient semantic search.
• Qdrant Vector Storage: Leverages Qdrant for scalable and fast vector similarity searches, forming the backbone of the RAG system.
• Intelligent Q&A Capability: Ask natural language questions about your documents and get contextual answers powered by Mistral AI's chat model.
• Local Data Focus: Keeps your data processing localized, ideal for sensitive or proprietary information.
• Real-time Knowledge Base: Ensures your Q&A agent always has the latest information as your documents change.

• An n8n instance (Cloud or self-hosted).
• Mistral AI API Key.
• Qdrant instance accessible by n8n (e.g., local Docker, cloud).
• A local directory on the n8n host machine or a mounted volume that n8n can access for file monitoring.

1. Download the n8n workflow JSON file.
2. Import the workflow into your n8n instance.
3. Configure the 'Local File Trigger' node: Set the Path parameter to the directory you want to monitor (e.g., /data/mydocuments). Ensure n8n has read/write access to this path if using Docker mounts.
4. In the 'Set Variables' node, update the directory variable if it's different from the trigger path and set your desired qdrant_collection name (e.g., local_file_search).
5. Configure Mistral AI Credentials: For 'Embeddings Mistral Cloud' and 'Mistral Cloud Chat Model' nodes, select or create your Mistral Cloud API credentials.
6. Configure Qdrant Credentials & Connection: For all 'Qdrant' nodes (HTTP Request nodes for delete/search and Qdrant Vector Store nodes), ensure the URL points to your Qdrant instance (e.g., http://qdrant:6333 if running Qdrant as a Docker container named 'qdrant' on the same network as n8n). Set up Qdrant API credentials if required by your Qdrant setup.
7. Verify the Qdrant collection name in the 'Qdrant Vector Store1' node (connected to the 'Vector Store Retriever') matches the one set in 'Set Variables' and used for insertions.
8. (Optional) Adjust text splitting parameters in the 'Recursive Character Text Splitter' node if needed for your document types.
9. Activate the workflow. New files in the monitored folder will be processed. Use the 'Chat Trigger' test URL or connect it to a chat interface to ask questions.