LPG Connector MCP Server

LPG Connector MCP Server

Enables Claude to query live LPG price and stock data, find nearest stations, view market trends, and submit crowdsourced price reports via MCP tools.

Category
Visit Server

README

LPG Connector — MCP Server

A remote MCP server that exposes live LPG price/stock data to Claude. Backed by in-memory mock data for now; built to swap to MongoDB with no changes to the tool logic.

Tools

Tool What it does
get_lpg_price(region?) Latest price/kg + stock per station in a region (omit region for all)
find_nearest_station(lat, lng, limit?) Nearest stations to a coordinate, with distance + latest price
get_market_trends(region, days?) Daily average price trend over the last N days
report_price(stationId, pricePerKg, stock, reporter?) Submit a crowdsourced price/stock report

stock is one of in_stock | low | out_of_stock.

Run locally

npm install
npm run dev        # tsx watch, http://localhost:3000/mcp

Health check: GET http://localhost:3000/health

Quick smoke test (list tools):

curl -s -X POST http://localhost:3000/mcp \
  -H "Content-Type: application/json" \
  -H "Accept: application/json, text/event-stream" \
  -d '{"jsonrpc":"2.0","id":1,"method":"tools/list","params":{}}'

Build for production

npm run build      # -> dist/
npm start          # node dist/index.js

Project layout

src/
  index.ts              Express + streamable HTTP transport (the /mcp endpoint)
  server.ts             MCP server + the 4 tool definitions
  store/
    types.ts            Domain types (Station, PriceReport, ...)
    DataStore.ts        Interface every store implements
    InMemoryStore.ts    JSON-seeded implementation (current)
  data/seed.json        Mock stations + price history

Swapping to MongoDB later

The tools only ever talk to the DataStore interface, so the migration is localized:

  1. npm install mongodb
  2. Create src/store/MongoStore.ts implementing DataStore (same method signatures as InMemoryStore). Map stations and priceReports to two collections.
  3. In src/index.ts, change one line:
    // const store = new InMemoryStore();
    const store = new MongoStore(process.env.MONGODB_URI!);
    

Nothing in server.ts changes. Seed the collections from data/seed.json once for a matching demo dataset.

Deploy (for the Claude custom connector)

Claude's cloud connects to this server over the public internet — a localhost URL won't work for claude.ai. This repo is set up for Render.

Deploy to Render

Option A — Blueprint (uses render.yaml, one click):

  1. Push this project to a GitHub repo.
  2. In Render: New → Blueprint, pick the repo. It reads render.yaml (build npm install && npm run build, start npm start, health check /health).
  3. Deploy. You'll get a URL like https://cylindr.onrender.com.

Option B — Manual web service:

  1. New → Web Service, connect the repo, set Root Directory to zserver.
  2. Build command: npm install && npm run build
  3. Start command: npm start
  4. Health check path: /health

After it's live

  • Confirm health: open https://<your-app>.onrender.com/health → {"status":"ok"}
  • Add to Claude: Settings → Connectors → Add custom connector, paste https://<your-app>.onrender.com/mcp, then enable it in a chat via the + menu.

Free-tier note: Render's free web services sleep after ~15 min idle and take ~30–60s to wake on the next request. Before your live demo, hit /health once to wake it so the first judge query isn't slow. Also: in-memory data is per-process, so a restart (or wake from sleep) resets to the seed and any report_price submissions are lost. The MongoDB swap removes both concerns.

Recommended Servers

playwright-mcp

playwright-mcp

A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.

Official
Featured
TypeScript
Magic Component Platform (MCP)

Magic Component Platform (MCP)

An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.

Official
Featured
Local
TypeScript
Audiense Insights MCP Server

Audiense Insights MCP Server

Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.

Official
Featured
Local
TypeScript
VeyraX MCP

VeyraX MCP

Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.

Official
Featured
Local
graphlit-mcp-server

graphlit-mcp-server

The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.

Official
Featured
TypeScript
Kagi MCP Server

Kagi MCP Server

An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.

Official
Featured
Python
E2B

E2B

Using MCP to run code via e2b.

Official
Featured
Neon Database

Neon Database

MCP server for interacting with Neon Management API and databases

Official
Featured
Exa Search

Exa Search

A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.

Official
Featured
Qdrant Server

Qdrant Server

This repository is an example of how to create a MCP server for Qdrant, a vector search engine.

Official
Featured