I spent about two weeks building an intelligent email management system. The idea came from seeing someone in my WeChat community share their company’s need for intelligent email processing and their search for solution partners, so I decided to build it by myself.
The main modules include:
- Gmail and MSN email integration (currently focused on reading new emails)
- Email formatting and processing,
- AI large model-based intelligent understanding of emails, such as content summarization, attachment summarization, event handling (e.g., meeting invitations), email priority classification, sentiment analysis, etc.
- Processed data is written into a database, managed by QuestDb.
Since email processing can take a long time, I used FastAPI to provide the service and Celery to manage the task queues. The API includes endpoints to check the task status. This allows third parties to call the API to check whether tasks have completed and whether the processed emails have been written to the database.
As a simple demo, I built a lightweight interface using Streamlit. System code was assisted with the help of AI coding tools (Claude, Qwen, DeepSeek, Gemini, ChatGPT, free versions). AI coder assist completing the basic framework, which is the base where I work on. Actually AI assit code significantly boosts efficiency for further development and refinement.
-
Previous
Top Python Libraries for Fintech and Trading -
Next
ML strategy optimisation and stock trading system
