StockPulse

image

An AI-powered Indian Stock Intelligence platform built entirely on Notion. Features a Python data pipeline that fetches daily price and delivery data from the NSE and BSE, screens 5,000+ stocks through a fundamental screener, and uses the Model Context Protocol (MCP) to allow AI assistants to read screened data, analyze fundamentals, and write comprehensive research reports directly into Notion.

View on GitHub

Tech Stack

  • Languages: Python
  • Technologies: Model Context Protocol (MCP), FastMCP, Notion API, NSE/BSE Data

Features

  • Data pipeline for downloading BhavCopy and delivery data from NSE/BSE.
  • 12-condition fundamental screener separating stocks by profitability, growth, governance.
  • Notion integration acting as a single source of truth across 5 linked databases.
  • Custom MCP server providing tools for AI models to query screened stocks and write reports.

Challenges and Solutions

  • Challenge: Handling scale of 5,000+ stocks concurrently.

    Solution: Built optimized and parallelized data pipelines from NSE and BSE directly into Notion.

Learning Outcomes

  • Gained deep experience in building custom MCP servers.
  • Architected Notion databases as a centralized UI/CMS for AI products.
  • Learned to design specific tool-calling capabilities that allow LLMs to dynamically reason.