mcp-servers/DEPLOYMENT_GUIDE.md

# MCP Job Search Stack - Deployment Guide

This guide walks you through setting up and deploying your two new MCPs for job searching and RFP discovery.

## Overview

You now have two complementary MCPs:

1. **rfp-scraper-mcp-server** - Discovers RFPs and freelance opportunities for control room and broadcast studio projects
2. **linkedin-mcp-server** - Searches LinkedIn jobs and manages professional networking

## Quick Start

### 1. Build Both MCPs

```bash
# Build RFP Scraper MCP
cd /sessions/focused-busy-tesla/mnt/MCP\ Servers/rfp-scraper-mcp-server
npm install
npm run build

# Build LinkedIn MCP
cd /sessions/focused-busy-tesla/mnt/MCP\ Servers/linkedin-mcp-server
npm install
npm run build
```

### 2. Verify Build Success

Both should complete without errors and create a `dist/` directory with compiled JavaScript.

```bash
# Test RFP Scraper
node rfp-scraper-mcp-server/dist/index.js --version

# Test LinkedIn
node linkedin-mcp-server/dist/index.js --version
```

## Integration with mcp-gateway

Your MCPs are designed to integrate with your existing mcp-gateway at `mcp.wilddragon.net`.

### Configuration

Add to your mcp-gateway configuration:

```json
{
  "mcpServers": {
    "rfp-scraper": {
      "command": "node",
      "args": ["/path/to/rfp-scraper-mcp-server/dist/index.js"],
      "env": {
        "RFP_SCRAPER_API_KEY": "${RFP_SCRAPER_API_KEY}"
      }
    },
    "linkedin": {
      "command": "node",
      "args": ["/path/to/linkedin-mcp-server/dist/index.js"],
      "env": {
        "LINKEDIN_ACCESS_TOKEN": "${LINKEDIN_ACCESS_TOKEN}"
      }
    }
  }
}
```

### Environment Variables

**For RFP Scraper:**
```bash
export RFP_SCRAPER_API_KEY=your_api_key  # Optional, for enhanced features
```

**For LinkedIn:**
```bash
export LINKEDIN_ACCESS_TOKEN=your_access_token  # Required for connection requests
```

Get your LinkedIn access token from: https://linkedin.com/developers

## Features by MCP

### RFP Scraper MCP (`rfp-scraper-mcp-server`)

Tools available:
- **rfp_search_projects** - Search for RFP opportunities
- **rfp_get_project_details** - Get detailed project information
- **rfp_filter_by_budget** - Filter projects by budget range

**Example Usage:**
```
Search for control room projects:
  keywords: ["control room", "broadcast studio"]
  limit: 20

Find high-budget opportunities:
  minBudget: 10000
  maxBudget: 50000
```

### LinkedIn MCP (`linkedin-mcp-server`)

Tools available:
- **linkedin_search_jobs** - Search for jobs
- **linkedin_get_job_details** - Get job information
- **linkedin_search_profiles** - Find professionals
- **linkedin_send_connection** - Connect with professionals
- **linkedin_get_profile** - View profiles

**Example Usage:**
```
Search for broadcast jobs:
  keywords: "broadcast engineer"
  jobType: "FREELANCE"

Find industry professionals:
  keywords: "broadcast systems"
  industry: "Broadcasting"
```

## Development Mode

For local testing and development:

```bash
# Terminal 1: RFP Scraper
cd rfp-scraper-mcp-server
npm run dev

# Terminal 2: LinkedIn
cd linkedin-mcp-server
npm run dev
```

This uses `tsx` to watch for changes and auto-reload.

## Testing the MCPs

### Using MCP Inspector

Test individual MCPs with the MCP Inspector:

```bash
# Test RFP Scraper
npx @modelcontextprotocol/inspector node rfp-scraper-mcp-server/dist/index.js

# Test LinkedIn
npx @modelcontextprotocol/inspector node linkedin-mcp-server/dist/index.js
```

The inspector provides a web interface to:
- List available tools
- Test tool calls
- View responses
- Debug issues

### Manual Testing

```bash
# Build both
npm run build

# Run and test
node rfp-scraper-mcp-server/dist/index.js

# In another terminal, send test request via stdin
echo '{"jsonrpc":"2.0","id":1,"method":"resources/list"}' | node rfp-scraper-mcp-server/dist/index.js
```

## Architecture Notes

### RFP Scraper
- **Service**: `services/scraper.ts` - Core scraping and filtering logic
- **Transport**: stdio (local) - designed for mcp-gateway integration
- **Error Handling**: Graceful handling of API errors and timeouts
- **Pagination**: Supports offset-based pagination for large result sets

### LinkedIn
- **Service**: `services/linkedin-client.ts` - LinkedIn API client
- **Transport**: stdio (local) - designed for mcp-gateway integration
- **Mock Data**: Currently uses mock data for development. Replace with real API calls once authenticated.
- **Authentication**: Supports LinkedIn OAuth tokens

## Production Deployment

### Prerequisites
- Node.js 18+
- npm or yarn
- Valid API credentials (LinkedIn token, etc.)

### Deployment Steps

1. **Install dependencies on production server:**
   ```bash
   npm install --production
   npm run build
   ```

2. **Set environment variables:**
   ```bash
   export LINKEDIN_ACCESS_TOKEN=prod_token
   export RFP_SCRAPER_API_KEY=prod_key
   ```

3. **Run MCPs via process manager (PM2):**
   ```bash
   pm2 start rfp-scraper-mcp-server/dist/index.js --name rfp-scraper
   pm2 start linkedin-mcp-server/dist/index.js --name linkedin
   ```

4. **Monitor:**
   ```bash
   pm2 logs
   pm2 status
   ```

## Troubleshooting

### Build Fails
- Ensure Node.js 18+ is installed: `node --version`
- Clear npm cache: `npm cache clean --force`
- Rebuild: `rm -rf node_modules dist && npm install && npm run build`

### Tools Not Appearing
- Check server started successfully: `npm start`
- Verify no TypeScript errors: `npm run build`
- Check tool names match exactly in your calls

### API Errors
- Verify environment variables are set: `echo $LINKEDIN_ACCESS_TOKEN`
- Check API credentials are valid
- Review error messages for guidance on next steps

### Connection Issues
- Ensure mcp-gateway can reach the servers
- Check firewall rules if using remote deployment
- Verify stdio transport is properly configured

## Next Steps

1. **Authenticate with APIs:**
   - Get LinkedIn access token: https://linkedin.com/developers
   - Set up any RFP API keys if needed

2. **Test Integration:**
   - Run MCPs locally first
   - Test with MCP Inspector
   - Verify with mcp-gateway

3. **Customize:**
   - Modify scraper keywords and sources as needed
   - Adjust job search filters
   - Add additional tools based on your workflow

4. **Monitor Usage:**
   - Track API rate limits
   - Monitor performance metrics
   - Adjust pagination limits if needed

## File Structure

```
MCP Servers/
├── rfp-scraper-mcp-server/
│   ├── src/
│   │   ├── index.ts
│   │   ├── types.ts
│   │   ├── constants.ts
│   │   └── services/
│   │       └── scraper.ts
│   ├── dist/
│   │   └── index.js (generated)
│   ├── package.json
│   ├── tsconfig.json
│   └── README.md
├── linkedin-mcp-server/
│   ├── src/
│   │   ├── index.ts
│   │   ├── types.ts
│   │   ├── constants.ts
│   │   └── services/
│   │       └── linkedin-client.ts
│   ├── dist/
│   │   └── index.js (generated)
│   ├── package.json
│   ├── tsconfig.json
│   └── README.md
└── DEPLOYMENT_GUIDE.md (this file)
```

## Support

For issues or questions:
1. Check individual README files in each MCP directory
2. Review error messages - they include suggested next steps
3. Check TypeScript compilation for type errors
4. Verify environment variables are set correctly

---

**Last Updated**: 2026-03-31
**Version**: 1.0.0