BLS MCP Server

# 🤖 Connecting Your BLS MCP Server to ChatGPT & AI Assistants ## 🌐 **Your Server Status** - **Public URL**: `https://uncriticisable-multilaterally-semaj.ngrok-free.dev` - **MCP Endpoint**: `https://uncriticisable-multilaterally-semaj.ngrok-free.dev/mcp` - **Status**: ✅ **ONLINE** (I can see traffic in your logs!) ## 🎯 **Method 1: ChatGPT Integration (Recommended)** ### **Step 1: Enable Developer Mode** 1. Open ChatGPT in your browser 2. Go to **Settings** → **Advanced Settings** 3. Toggle on **Developer Mode** ### **Step 2: Create MCP Connector** 1. Go to **Settings** → **Connectors** → **Create** 2. Fill in the details: - **Connector Name**: `BLS Data Server` - **Description**: `Access Bureau of Labor Statistics data including CPI, employment, and economic indicators` - **Connector URL**: `https://uncriticisable-multilaterally-semaj.ngrok-free.dev/mcp` - **Authentication**: `None` (for now) 3. Click **Create** ### **Step 3: Use in ChatGPT** 1. Start a new chat 2. Click the **+** button near the message composer 3. Select **Developer Mode** 4. Toggle on your **BLS Data Server** connector 5. Test with prompts like: - "Use the BLS Data Server to get CPI data for all items" - "List available economic data series" - "Show me food price inflation data" ## 🔧 **Method 2: Direct API Integration** ### **For Custom Applications** ```python import requests import json # Your MCP server endpoint MCP_URL = "https://uncriticisable-multilaterally-semaj.ngrok-free.dev/mcp" # Example: List available tools def list_tools(): response = requests.post(MCP_URL, json={ "jsonrpc": "2.0", "id": 1, "method": "tools/list", "params": {} }) return response.json() # Example: Get CPI data def get_cpi_data(): response = requests.post(MCP_URL, json={ "jsonrpc": "2.0", "id": 2, "method": "tools/call", "params": { "name": "get_series", "arguments": { "series_id": "CUUR0000SA0", "start_year": 2023, "end_year": 2024 } } }) return response.json() # Test the connection print("Available tools:", list_tools()) print("CPI data:", get_cpi_data()) ``` ## 🧪 **Method 3: Test Your Integration** ### **Test Commands** ```bash # Test health endpoint curl https://uncriticisable-multilaterally-semaj.ngrok-free.dev/health # Test MCP endpoint curl -X POST https://uncriticisable-multilaterally-semaj.ngrok-free.dev/mcp \ -H "Content-Type: application/json" \ -d '{ "jsonrpc": "2.0", "id": 1, "method": "tools/list", "params": {} }' # Test SSE stream curl -N https://uncriticisable-multilaterally-semaj.ngrok-free.dev/sse ``` ## 📋 **Available Tools for AI Assistants** Your server provides these tools that AI can use: ### **1. `get_series`** - **Purpose**: Fetch BLS data series by ID - **Example**: Get CPI data for all items - **Parameters**: `series_id`, `start_year`, `end_year` ### **2. `list_series`** - **Purpose**: List available BLS data series - **Example**: Show all CPI-related series - **Parameters**: `category`, `limit` ### **3. `get_series_info`** - **Purpose**: Get detailed metadata about a series - **Example**: Get description and availability info - **Parameters**: `series_id` ## 🎯 **Sample Data Available** | Series ID | Description | Category | |-----------|-------------|----------| | CUUR0000SA0 | CPI All Items | CPI | | CUUR0000SAF | CPI Food | CPI | | CUUR0000SAF11 | CPI Food at Home | CPI | | CUUR0000SETA | CPI Energy | CPI | | CUUR0000SAH | CPI Housing | CPI | | CUUR0000SAT | CPI Transportation | CPI | | CUUR0000SAM | CPI Medical Care | CPI | ## 🔄 **Method 4: Claude Desktop (Alternative)** If you prefer Claude Desktop: 1. Edit `~/Library/Application Support/Claude/claude_desktop_config.json`: ```json { "mcpServers": { "bls-data": { "command": "uv", "args": [ "run", "python", "/Users/rafaelkovashikawa/Downloads/projects/bls_food/bls_mcp/scripts/start_server.py" ], "cwd": "/Users/rafaelkovashikawa/Downloads/projects/bls_food/bls_mcp" } } } ``` 2. Restart Claude Desktop ## 🚀 **Quick Start for ChatGPT** **Right now, you can:** 1. **Open ChatGPT** → Settings → Developer Mode (ON) 2. **Create Connector**: - Name: `BLS Economic Data` - URL: `https://uncriticisable-multilaterally-semaj.ngrok-free.dev/mcp` 3. **Start chatting**: - "Use BLS Economic Data to get current inflation rates" - "Show me food price trends from the BLS server" - "What economic indicators are available?" ## 🔧 **Troubleshooting** ### **If ChatGPT can't connect:** 1. **Check server status**: `curl https://uncriticisable-multilaterally-semaj.ngrok-free.dev/health` 2. **Verify MCP endpoint**: Test the `/mcp` endpoint directly 3. **Check ngrok tunnel**: Make sure it's still active ### **If you get 404 errors:** - The server is running but some endpoints return 404 (this is normal) - Use the specific endpoints: `/health`, `/mcp`, `/sse` ## 🎉 **You're Ready!** Your BLS MCP server is **live and accessible** to ChatGPT and other AI assistants. The traffic logs show it's already being accessed from various IP addresses, which means it's working perfectly! **Next step**: Try the ChatGPT integration above! 🚀