AI Document Processor & Note Generation Agent (Mistral & Qdrant)

Unlock Automated Content Understanding & Generation with this AI Agent

This n8n workflow acts as a powerful AI Agent designed to streamline your document processing and content creation tasks. It watches a specified local folder for new documents (PDF, DOCX, or plain text files). Upon detecting a new file, the agent extracts its content, generates an initial AI-powered summary, and then intelligently breaks down the content, creating vector embeddings using Mistral AI and storing them in a Qdrant vector database.

This enables a sophisticated Retrieval Augmented Generation (RAG) process. For predefined document templates (e.g., Study Guide, Timeline, Briefing Document), the AI Agent first generates probing questions about the source document, then uses the vector store to retrieve the most relevant information to answer these questions. Finally, it synthesizes this information into a new, structured document in Markdown format, tailored to the chosen template, and saves it to your filesystem.

This AI Agent excels at transforming raw information into organized, actionable knowledge, significantly boosting productivity for research, content repurposing, and knowledge management.

Key Features & Benefits

• Automated Document Ingestion: Monitors a local folder and processes new PDF, DOCX, and TXT files automatically.
• AI-Powered Summarization: Utilizes Mistral AI (e.g., open-mixtral-8x7b) via Langchain to generate concise summaries of ingested documents.
• Intelligent Vectorization: Creates vector embeddings of document content using Mistral Embeddings and stores them in a Qdrant vector database for efficient similarity search.
• Advanced RAG Pipeline: Employs a multi-step Retrieval Augmented Generation process to generate contextually rich and accurate content.
• Multi-Format Note Generation: Capable of producing various structured outputs like Study Guides, Timelines, and Briefing Documents based on customizable templates and AI-driven analysis.
• Content Repurposing: Easily transform long-form documents into various shorter, focused pieces of content.
• Automated Export: Saves the generated notes directly to your local filesystem, organized alongside the original source material.
• Customizable Templates: Define your own document types and generation instructions for tailored outputs.

• An n8n instance (Cloud or self-hosted).
• Mistral AI API Key with access to suitable models (e.g., open-mixtral-8x7b).
• Qdrant instance (self-hosted or cloud) URL and API key if required.
• A designated local folder path for the 'Local File Trigger' node to monitor and for exporting generated files.
• The @n8n/n8n-nodes-langchain node package installed in your n8n instance.

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 absolute path of the folder you want this agent to monitor for new documents (e.g., /home/node/storynotes/context).
4. Configure Mistral AI Credentials: In all 'Embeddings Mistral Cloud' and 'Mistral Cloud Chat Model' nodes, select or create your Mistral AI API credentials.
5. Configure Qdrant Connections: In both 'Qdrant Vector Store' nodes (one for insert, one for retrieval via 'Vector Store Retriever'), configure the 'Qdrant API' credentials with your Qdrant instance URL and API key (if secured). Ensure the 'Qdrant Collection' name (e.g., 'storynotes') is consistent.
6. Customize Document Templates (Optional): Modify the 'Get Doc Types' node to add, remove, or change the document templates (filename, title, description). The descriptions are used in prompts for the AI.
7. Review Generation Prompts: Check the system messages in the 'Interview' and 'Generate' LLM Chain nodes to fine-tune how the AI generates questions and final documents.
8. Verify File Paths: Ensure the path logic in the 'Export to Folder' node (fileName parameter) correctly constructs output paths based on your system and the 'Local File Trigger' base path.
9. Activate the workflow. Add new .pdf, .docx, or .txt files to the monitored folder to trigger the AI agent and generate your notes.