Instantly Search Emails with AI: Hybrid RAG Chatbot for Telegram
Retrieve accurate email information in seconds, reducing manual search time by up to 90% and improving response efficiency.
Manually sifting through extensive email archives for specific information consumes valuable time and hinders productivity. This workflow deploys an advanced AI chatbot, accessible via Telegram, that provides instant and precise answers to your email queries by intelligently combining semantic and structured search.
Documentation
AI-Powered Email Chatbot with Hybrid RAG
Navigate your personal email archives with unprecedented ease. This n8n workflow creates a sophisticated AI chatbot that responds to your natural language queries, delivering precise information from your emails by leveraging both semantic (vector embeddings) and structured (SQL database) search capabilities. Engage with it conveniently via Telegram or directly within n8n's chat interface.
Key Features
- Dual RAG Approach: Combines vector embeddings for semantic content understanding with structured SQL queries for precise data points, ensuring comprehensive and accurate email retrieval.
- Natural Language Interaction: Ask questions about your emails in plain English and receive intelligent, context-aware responses, eliminating the need for complex database queries.
- Telegram Integration: Interact with your AI email assistant directly from your Telegram chat, enabling convenient and efficient on-the-go access to your email archive.
- Conversational Memory: The chatbot maintains context across interactions, providing a more fluid, personalized, and intelligent conversational experience.
- Automated Response Formatting: Handles potentially large responses by splitting them into manageable chunks and correctly formatting them for Telegram's MarkdownV2, ensuring readability.
How It Works
The workflow activates upon receiving messages from either a configured Telegram bot or n8n's internal chat interface. Your natural language query is then directed to an advanced AI agent, powered by LangChain. This agent intelligently determines the best retrieval strategy, performing either a semantic search using Pgvector (for understanding the context and meaning of email content) or a structured SQL query (for precise data points like dates, senders, or specific fields) against your email database. The agent can combine results from both methods to formulate a comprehensive answer. The generated response is then automatically beautified, split into smaller chunks if lengthy, escaped for Markdown compatibility, and delivered back to you via Telegram or the n8n chat, providing rapid and accurate access to your email information.