home / mcp / gullakai mcp server

GullakAI MCP Server

A lightweight MCP server providing finance Q&A, news simplification, expense tracking, and city purchasing power comparisons.

Installation

Add the following to your MCP client configuration file.

Configuration

View docs

{
  "mcpServers": {
    "anjulbhatia-gullak-mcp": {
      "command": "python",
      "args": [
        "server.py"
      ],
      "env": {
        "MY_NUMBER": "your_whatsapp_number_here",
        "AUTH_TOKEN": "your_secure_token_here"
      }
    }
  }
}

GullakAI MCP Server enables a personal finance assistant that you can run locally to manage budgets, track expenses, summarize financial news, and compare purchasing power across Indian cities. It provides easy, multilingual finance Q&A, a simple expense/budget workflow, and city-level purchasing power checks that you can access from a compatible MCP client like WhatsApp or a browser chatbot.

How to use

Interact with GullakAI MCP Server through your MCP client to ask finance questions, get plain-language explanations of financial news, and manage your expenses and budgets via simple chat commands. You can also compare cost-of-living and local purchasing power across cities.

How to install

Follow these concrete steps to install and run GullakAI MCP Server on your machine. These steps assume you are comfortable using a terminal and editing environment variables.

uv .venv
uv sync
source .venv/activate

Create a configuration file with your authentication token and the number you will use to connect the MCP client.

AUTH_TOKEN=your_secure_token_here
MY_NUMBER=your_whatsapp_number_here

Add purchasing power data by creating a file named purchasingpower.json in the root folder with a structure like this (include more cities as needed).

{
  "cities": [
    {"city": "Hyderabad, India", "local_purchasing_power_index": 154.1},
    {"city": "Bangalore, India", "local_purchasing_power_index": 149.7}
  ]
}

Finally, run the MCP server using the command below.

python server.py

Additional notes and configuration

Expense and budget data are cached in memory with a 7-day expiration. This setup is designed for asynchronous operation and works well with both WhatsApp and browser-based MCP clients.

The server supports a few practical expense commands you can use in your chat: set budget <month> <category> <amount>, spent <amount> on <category>, owe <person/description> <amount>, bill <description> <amount> due YYYY-MM-DD, and summary [month]. You can also ask natural-language questions about personal finances or run purchasing power comparisons.

License

This project is licensed under the MIT License.

Available tools

Finance Q&A

Conversational finance assistant that answers personal finance questions in multiple languages.

News Simplifier

Summarizes and explains financial news and jargon in easy-to-understand language.

Expense & Budget Monitor

Manage budgets, record expenses, track debts and bills through simple chat commands.

Purchasing Power Checker

Compare affordability across cities using local purchasing power data.