SIGNALS MCP Server — Market Readiness Index

MCP (Model Context Protocol) server exposing all 7 SIGNALS market readiness signals as tools. Runs server-side — no CORS issues. All Yahoo Finance, ENTSOG, AGSI+, Carbon Intensity, and Open-Meteo APIs work cleanly.

Tools Exposed

Tool	Signal	Source
get_norwegian_gas_flows	S1 — Norwegian Gas Flows	ENTSOG + UMMs
get_eu_gas_storage	S2 — EU Gas Storage	AGSI+
get_nbp_gas_price	S3 — NBP Gas Price	Yahoo (TTF)
get_uk_generation_mix	S4 — UK Gen Mix	Carbon Intensity
get_temperature_vs_norm	S5 — Temperature	Open-Meteo
get_implied_power_price	S6 — Implied Power	Computed
get_brent_crude	S7 — Brent Crude	Yahoo (BZ=F)
get_market_readiness_index	ALL 7 + MRI	All sources

Quick Start (Local)

pip install -r requirements.txt

# STDIO transport (for Claude Desktop, Claude Code)
python signals_mcp.py

# SSE transport (for Base44, remote clients)
python signals_mcp_sse.py

Deploy to Railway (Recommended)

Railway is the easiest way to get a hosted URL for Base44.

1. Push this folder to a GitHub repo
2. Go to railway.app → New Project → Deploy from GitHub
3. Railway auto-detects the Dockerfile
4. Once deployed, your SSE endpoint is: https://your-app.railway.app/sse

Alternative: Render.com

1. New Web Service → connect your repo
2. Build Command: pip install -r requirements.txt
3. Start Command: python signals_mcp_sse.py
4. SSE endpoint: https://your-app.onrender.com/sse

Connect to Base44

1. Go to Account Settings → MCP Connections
2. Click Add MCP Server
3. Enter:
   • Name: SIGNALS Market Readiness
   • URL: https://your-deployed-url.railway.app/sse
4. Save

Now in the Base44 AI builder chat, you can say things like:

• "Pull the current Market Readiness Index from SIGNALS"
• "Check Norwegian gas flows"
• "What's the implied UK power price?"

Note: Base44 MCP connections work in the AI builder chat, not in deployed apps. For deployed apps, use Base44 backend functions that call your server's tools via HTTP, or use the Custom OpenAPI integration approach (see below).

Connect to Claude Desktop

Add to claude_desktop_config.json:

{
  "mcpServers": {
    "signals": {
      "command": "python",
      "args": ["/path/to/signals_mcp.py"]
    }
  }
}

Connect to Claude Code

claude mcp add signals python /path/to/signals_mcp.py

For Base44 Deployed Apps: Backend Functions

If you want live signal data in your actual Base44 app (not just builder chat), create backend functions that call your hosted server directly:

// Base44 backend function example
import fetch from 'node-fetch';

export async function getMarketReadiness() {
  // Call your hosted MCP server's underlying endpoint
  const response = await fetch('https://your-app.railway.app/sse', {
    method: 'POST',
    headers: { 'Content-Type': 'application/json' },
    body: JSON.stringify({
      method: 'tools/call',
      params: { name: 'get_market_readiness_index', arguments: {} }
    })
  });
  return await response.json();
}

Alternatively, add a simple REST wrapper (see rest_wrapper.py if included) that exposes /api/signals as a standard JSON endpoint your Base44 app can call.

Updating the Power Baseline

Signal 6 uses a baseline from smart-energy.uk. Update POWER_BASELINE in signals_mcp.py, or pass it as a parameter when calling get_implied_power_price(baseline_gbp_mwh=72.50).

Architecture

┌─────────────┐     ┌──────────────────┐     ┌──────────────┐
│  Base44 AI  │────▶│  SIGNALS MCP     │────▶│  ENTSOG      │
│  Builder    │ SSE │  Server          │     │  AGSI+       │
│  Chat       │     │  (Railway/Render)│     │  Yahoo       │
├─────────────┤     │                  │     │  CarbonInt   │
│  Claude     │────▶│  Python + MCP SDK│     │  Open-Meteo  │
│  Desktop    │stdio│                  │     └──────────────┘
├─────────────┤     └──────────────────┘
│  Base44 App │────▶ (via backend function / REST wrapper)
│  (deployed) │ HTTP
└─────────────┘