GWM MCP Server

MCP (Model Context Protocol) server for Generic Web Monitor. Interact with your GWM competitive intelligence instance via Claude Desktop or Claude Code.

Features

• Query Data: Companies, URLs, changes, analysis results, market metrics
• Modify Configuration: Add companies/URLs, update settings, manage saved views
• Trigger Workflows: Run scrapes, analysis, deployments directly from Claude

Quick Start

1. Get Your Connection String

From Heroku:

heroku config:get DATABASE_URL -a your-app

From Neon Dashboard: Connection Details → Connection string

2. Configure Claude Desktop

Open ~/Library/Application Support/Claude/claude_desktop_config.json (macOS) or %APPDATA%\Claude\claude_desktop_config.json (Windows):

{
  "mcpServers": {
    "gwm": {
      "command": "npx",
      "args": ["-y", "gwm-mcp"],
      "env": {
        "POSTGRES_CONNECTION_STRING": "postgres://user:pass@host:5432/db?sslmode=require"
      }
    }
  }
}

3. Restart Claude Desktop

The GWM tools should now appear. Try: "List all companies in my GWM instance"

Configuration

Required

| Variable | Description | |----------|-------------| | POSTGRES_CONNECTION_STRING | PostgreSQL connection string |

Optional (for workflow triggers)

| Variable | Description | |----------|-------------| | GITHUB_TOKEN | GitHub personal access token with workflow scope | | GITHUB_OWNER | GitHub organization/user (e.g., "Redmore-Studio") | | GITHUB_REPO | Repository name (e.g., "gwm-test") |

Optional (for AI discovery)

| Variable | Description | |----------|-------------| | PERPLEXITY_API_KEY | Perplexity API key for URL/market discovery |

Available Tools

Read Operations

| Tool | Description | |------|-------------| | gwm_list_companies | List monitored companies | | gwm_get_company | Get company details | | gwm_list_urls | List monitored URLs | | gwm_get_url_status | Get URL scrape status | | gwm_list_changes | List detected changes | | gwm_get_change_details | Get change with diff | | gwm_get_baseline_analysis | Get entity extraction | | gwm_get_enhanced_analysis | Get deep analysis | | gwm_list_market_metrics | List market metrics | | gwm_get_competitor_facts | Get competitor facts | | gwm_search_kwic | Search keyword matches | | gwm_list_keywords | List tracked keywords | | gwm_list_entity_groups | List entity normalizations | | gwm_health_check | System health status |

Write Operations

| Tool | Description | |------|-------------| | gwm_add_company | Add company to monitor | | gwm_update_company | Update company settings | | gwm_add_url | Add URL to monitor | | gwm_update_url | Update URL settings | | gwm_disable_url | Disable URL from monitoring | | gwm_update_company_metric | Update company sizing | | gwm_create_saved_view | Create filter view | | gwm_update_saved_view | Update saved view | | gwm_delete_saved_view | Delete saved view |

Workflow Operations

| Tool | Description | |------|-------------| | gwm_trigger_full_monitor | Run full pipeline | | gwm_trigger_scrape | Run scraping only | | gwm_trigger_deploy | Quick deploy | | gwm_get_workflow_runs | Check workflow status |

Discovery Operations (requires PERPLEXITY_API_KEY)

| Tool | Description | |------|-------------| | gwm_discover_urls | Find URLs to monitor for a company using AI | | gwm_discover_market | Discover competitors and URLs for a market |

Example Usage

You: List all companies in the AI market

Claude: Here are the companies in the ai_companies market:
- OpenAI [FOCAL] (tier1)
  Category: AI Lab
  Website: https://openai.com
  URLs: 5
...

You: Add Mistral AI as a new competitor

Claude: Company Added!
Mistral AI (ID: 52)
- Website: https://mistral.ai
- Category: AI Lab
- Market: ai_companies

Run a scrape workflow to start monitoring.

You: Trigger a full monitor for the AI market

Claude: Full Monitor Triggered!
Status: Workflow triggered successfully
Run ID: 12345678

Pipeline typically takes 15-20 minutes...

Development

Build from source

cd mcp-server
npm install
npm run build

Run locally

POSTGRES_CONNECTION_STRING="postgres://..." node dist/index.js

Troubleshooting

"Connection timeout" error

Your database may be sleeping (cold start). Wait a few seconds and retry.

"Schema not found" error

The database may not be initialized. Run the "Initialize Database" workflow in GitHub Actions.

Workflow triggers not working

Ensure GITHUB_TOKEN, GITHUB_OWNER, and GITHUB_REPO are all set in your MCP config.

License

MIT