How do I use Maharashtra Medicine MCP Server?

1. Click on "Install Server". 2. Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state. 3. In the chat, type @ followed by the MCP server name and your instructions, e.g., "@Maharashtra Medicine MCP Server Which medicines are expiring within the next 30 days?" That's it! The server will respond to your query, and you can continue using it as needed. Here is a step-by-step guide with screenshots.

Maharashtra Medicine MCP Server — Cloudflare Workers

FastMCP server for Maharashtra wholesale medicine purchase analysis, deployed as a Python Worker on Cloudflare's global edge network.

Project Structure

medicine-mcp-server/
├── src/
│   └── worker.py                         ← FastMCP server (all 10 tools)
├── data/
│   └── maharashtra_wholesale_medicine_purchase.csv   ← upload to R2, not deployed
├── .github/
│   └── workflows/
│       └── deploy.yml                    ← GitHub Actions auto-deploy
├── wrangler.toml                         ← Cloudflare Workers config
├── pyproject.toml                        ← Python deps + pywrangler config
└── README.md

Architecture

MCP Client (Claude Desktop / mcp-remote)
        │  HTTPS  (Streamable HTTP transport)
        ▼
Cloudflare Workers Edge
  ┌─────────────────────────────┐
  │  on_fetch()                 │
  │     │                       │
  │     ▼                       │
  │  MedicineServer             │
  │  (Durable Object)           │
  │     │                       │
  │     ├─ FastMCP tools ──────►│ env.GEMINI_API_KEY  (Secret)
  │     │                       │
  │     └─ _get_df() ──────────►│ env.MEDICINE_DATA   (R2 bucket)
  └─────────────────────────────┘

CSV data lives in Cloudflare R2 — fetched once per Worker isolate, then cached in memory for the lifetime of the isolate.
Gemini API key is stored as a Cloudflare Secret — never in source code.
Transport is Streamable HTTP at /mcp — the current MCP spec standard.

One-Time Setup

Prerequisites

# Node.js 18+  (for wrangler CLI)
node --version

# Python 3.11+ and uv
pip install uv
uv tool install pywrangler

# Wrangler CLI
npm install -g wrangler

# Authenticate wrangler with your Cloudflare account
npx wrangler login

Step 1 — Clone & configure

git clone https://github.com/YOUR_ORG/medicine-mcp-server.git
cd medicine-mcp-server

Open wrangler.toml and uncomment + fill in your account_id:

account_id = "abc123..."    # from dash.cloudflare.com → Overview

Step 2 — Create the R2 bucket

npx wrangler r2 bucket create medicine-mcp-data

Step 3 — Upload the CSV to R2

npx wrangler r2 object put medicine-mcp-data/maharashtra_wholesale_medicine_purchase.csv \
  --file ./data/maharashtra_wholesale_medicine_purchase.csv

Verify the upload:

npx wrangler r2 object get medicine-mcp-data/maharashtra_wholesale_medicine_purchase.csv \
  --file /tmp/verify.csv && wc -l /tmp/verify.csv

Step 4 — Store your Gemini API key as a secret

npx wrangler secret put GEMINI_API_KEY
# Paste your key when prompted (get one free at https://aistudio.google.com/app/apikey)

Step 5 — Deploy

uv run pywrangler deploy

Your server is now live at:

https://medicine-mcp-server.<your-account>.workers.dev/mcp

GitHub Actions Auto-Deploy

Every push to main triggers an automatic deploy via .github/workflows/deploy.yml.

Add these two secrets in your GitHub repo (Settings → Secrets and variables → Actions):

Secret	Where to find it
`CLOUDFLARE_API_TOKEN`	dash.cloudflare.com → My Profile → API Tokens → Create Token → "Edit Cloudflare Workers" template
`CLOUDFLARE_ACCOUNT_ID`	dash.cloudflare.com → Overview → API section

After adding secrets, any git push origin main will auto-deploy.

Local Development

# Start local dev server (uses local R2 simulation — no real R2 calls)
uv run pywrangler dev

# Server available at http://localhost:8787/mcp
# Test with MCP Inspector: npx @modelcontextprotocol/inspector http://localhost:8787/mcp

Connecting Claude Desktop

Add to claude_desktop_config.json:

{
  "mcpServers": {
    "medicine": {
      "command": "npx",
      "args": [
        "mcp-remote",
        "https://medicine-mcp-server.<your-account>.workers.dev/mcp"
      ]
    }
  }
}

Available Tools

#	Tool	Purpose
1	`search_medicines`	Search by product / manufacturer / supplier / buyer
2	`get_invoice_details`	Full line-items for one or more invoices
3	`filter_by_schedule`	Filter by drug schedule (H, H1, X, G, OTC)
4	`get_expiry_alerts`	Medicines expiring within N days
5	`analyse_supplier`	Spend & invoice summary for a supplier
6	`analyse_buyer`	Purchase history & schedule mix for a buyer
7	`top_products_by_spend`	Ranked products by taxable amount / quantity
8	`gst_summary`	CGST / SGST / IGST breakdown by invoice/supplier/buyer
9	`cold_chain_and_narcotic_items`	Cold-chain & Schedule X items
10	`natural_language_query`	Free-form NL question answered by Gemini

Updating the CSV

To replace the CSV with a newer file:

npx wrangler r2 object put medicine-mcp-data/maharashtra_wholesale_medicine_purchase.csv \
  --file ./data/your_new_file.csv

# The Worker's in-memory cache resets automatically on the next isolate cold-start.
# Force an immediate reset by redeploying:
uv run pywrangler deploy

Useful Commands

# Stream live logs from production
uv run pywrangler tail

# List R2 objects
npx wrangler r2 object list medicine-mcp-data

# List secrets
npx wrangler secret list

# Delete a secret
npx wrangler secret delete GEMINI_API_KEY

Maharashtra Medicine MCP Server