Email Agent

Small learning project built to experiment with LangChain, Flask, the Gmail API and basic RAG.

The app can read recent Gmail threads, pass them to a LangChain agent, and generate email drafts/replies.
For the RAG part, I used a small fake CSV knowledge base about a fictional company called NexaTech.

Tech used

• Python
• Flask
• LangChain / LangGraph
• Gmail API
• FAISS
• OpenAI embeddings

Notes

This is not a production-ready email automation tool.
It was built mainly to learn how to combine agents, tool calling and retrieval-augmented generation.

Sensitive files like .env, credentials.json, token.json, email logs and local vector indexes are not included in the repository.