Spotlight Cash MCP server
This repository contains an implementation of an MCP (Model Context Protocol) Server, which allows seamless integration between external systems and AI assistants that support MCP.
Overview
This repository contains a suite of data services that work together to:
- Handle authenticated “session” creation and message flow (backed by an AWS SQS).
- Uses Ventura Securities' and Nifty Trader's data APIs for getting the financial data of a financial instrument.
- Provide an MCP-oriented financial instrument valuation service with login and notification listening.
It’s designed to be simple to run locally while remaining cloud-ready.
Capabilities
You can explore as much as you want with this MCP server.
- Forecast a stock's price based on GARP or DCF model evaluations
- Create hedge strategies that are delta neutral and more inclined towards your preference be it stock-based or index-based. If you are looking for covered calls or synthetic collar based strategies.
- Understand technicals, fundamentals, valuations and then trade scrips smartly.
- Identify chart patterns of 1Day, 1Week, 1Month, 1Year etc. for finding good technical setups.
- Analyse your trading patterns using your trade reports or understand charges in your ledger summary.
- Start and end your day by asking the market cues to stay updated around the financial markets of the world.
- You can ask opinions on your custom strategies and if they have a high probability of profit.
- Ask which stocks are approaching an indicator like golden cross or death cross
- See if any stocks from your portfolio require rotation based on some kind of indicator that you follow
- Ask suggestions on which IPOs to apply for? takes considerations from GMP present on chittorgarh.in
- Fetch volume boomers to gain an edge on trading NSE stocks that have unusually high volume.
and many more
Features
- Session service with queue integration (e.g., AWS SQS).
- Data service for financial data feeding to LLMs
- MCP service with SSO authentication and long-running notification listening.
- Clean separation of concerns so each service can be scaled independently.
Requirements
- Python 3.11.9
- virtualenv (for an isolated environment)
- OS: macOS, Linux, or Windows
- Possibly a CRON service to download the master files(bhavcopy) daily from your broker or a vendor to remain updated if any stock's name or nse, nfo or bse token has been changed
Quick start
1) Create and activate a virtual environment 2) Sync the packages from requirements.txt 3) You can either run the project using Docker Compose or you can run the server_data.py and server_mcp.py file on different ports