Chat with your email history using Telegram Mistral and pgvector for RAG

Say goodbye to forgotten emails and lost memories!

This automation gives you an AI-powered memory over your inbox — letting you ask questions like “Where did I stay last summer?” or “What were my grades?” and get real answers using semantic and structured email search.

Built entirely on open-source tools, this local-first solution is perfect for privacy-conscious power users and devs looking to explore vector-based search with AI.

What it does:

📥 Uses vector similarity search to semantically analyze your email content

🧠 Combines structured and semantic queries via natural language

🗄️ Runs entirely locally using Ollama (LLM), nomic-embed-text (embeddings), and PGVector (PostgreSQL)

🔄 Integrates with Telegram for a natural chat interface

✅ Setup guide & importable automation included

Need help setting it up? We offer full configuration and testing for a one-time fee.