Multi-language Metadata Generation Workflow

Name: Multi-language Metadata Generation Workflow
Rating: 5 (5 reviews)
Author: Free N8N

Intermediate

13 nodes connected

detail.loadingPreview

Free N8N Temples

220 views

13 downloads

AutomationAIGroqLLMPineconeRAGclassificationdocumentationemailmetadatamultilingual

This workflow automates the generation of structured metadata in both English and Chinese from incoming emails.

About This Workflow

This workflow is designed to process incoming emails, extract relevant information, classify them into predefined categories, and then generate structured metadata in both English and Chinese. It leverages AI models for classification and content generation, with a robust error handling mechanism and data enrichment capabilities.

Key Features

Email ingestion via IMAP.
Intelligent email classification using both code and AI.
AI-powered content generation for metadata (English and Chinese).
Knowledge base integration for enriched responses.
Conditional routing for different email categories (HR, billing, complaint, etc.).
Email sending for notifications and replies.
PDF file processing for data loading.

How To Use

Email Trigger: Configure the IMAP node to connect to your email server and fetch new emails.
Classification (Code & Switch): The Code node first attempts to classify emails based on keywords. The Switch node then routes emails based on the determined category.
AI Classification (Fallback): If the initial classification is insufficient, the Basic LLM Chain and Groq Chat Model nodes can be used as a more advanced AI-based classifier.
HR Workflow: For emails categorized as 'hr', the Basic LLM Chain and Groq Chat Model nodes are used to evaluate candidate suitability for a role.
Inquiry & Knowledge Base: For inquiries, the RAG INQURY REPLY node utilizes a Pinecone Vector Store and Embeddings Cohere to retrieve relevant information from a knowledge base and generate a response.
Metadata Generation (English & Chinese): The Metadata Generator (Code) node is the core component for creating the structured JSON output. It constructs both English and Chinese metadata fields, including titles, summaries, and markdown content.
Data Loading and Indexing: Nodes like HTTP Request, Extract from File, Pinecone Vector Store, Embeddings Cohere, Default Data Loader, and Recursive Character Text Splitter are used for ingesting and indexing data, particularly for the knowledge base.
Email Notifications: Various emailSend nodes are configured to notify relevant teams or reply to customers based on the email category and AI processing.

Apps Used

Groq

LLM

Pinecone

RAG

classification

documentation

metadata

multilingual

Workflow JSON

{
  "id": "af811344-8851-4bb3-8dc4-c39600437916",
  "name": "Multi-language Metadata Generation Workflow",
  "nodes": 13,
  "category": "Automation",
  "status": "active",
  "version": "1.0.0"
}

Note: This is a sample preview. The full workflow JSON contains node configurations, credentials placeholders, and execution logic.