MCP (Model Context Protocol)

What is MCP?

MCP (Model Context Protocol) is an open standard that lets AI assistants connect directly to the tools you use every day. With PitchGhost's MCP integration, you can ask your favorite AI assistant to create Ghosts, review Finds, trigger scans, and more — all without leaving your AI client.

Instead of switching between tabs, your AI assistant works with PitchGhost on your behalf, understands your business context, and helps you act on opportunities faster.

Features

Manage Your Ghosts

Create new Ghosts (What's a Ghost?), update their settings, enable or disable scanning, and trigger scans on demand. Your AI assistant has access to built-in guides for each platform's search syntax, so it can help you craft effective search queries for Twitter, LinkedIn, Reddit, and Facebook.

You can also delete Ghosts you no longer need (requires Admin or Owner permissions).

Review and Act on Finds

Search and filter your Finds across any Ghost or platform. Your AI assistant can read full Find details, including the original content, AI analysis, and confidence scores.

Take action on Finds by assigning them to team members, tagging them, setting result statuses, adding notes, providing feedback, or marking favorites — all through conversation.

Search Platforms in Real Time

Run live searches directly from your AI assistant. This is useful for exploring what's being said about a topic before creating a Ghost, or for one-off research that doesn't need ongoing monitoring.

Explore Your Account

Your AI assistant can look up your company details, team members, and connected integration accounts. This helps it understand your business context and make better suggestions.

Built-in Platform Guides

Your AI assistant automatically receives search syntax guides for every platform PitchGhost supports. When creating or editing a Ghost, it already knows the available operators, filters, and character limits for Twitter, LinkedIn, Reddit, and Facebook — so you don't have to.

Supported AI Clients

PitchGhost's MCP integration works with any AI client that supports the Model Context Protocol, including:

• Claude Desktop — Anthropic's desktop app

• Claude Code — Anthropic's CLI for developers

• ChatGPT — OpenAI's main chat app

• Cursor — AI-powered code editor

• VS Code — Via GitHub Copilot's MCP support

• Notion — AI-powered workspace

Connecting the Integration

Step 1: Open the MCP Setup Sheet

Navigate to Settings → Integrations in your PitchGhost account. Find the MCP integration and click Setup.

Step 2: Choose Your AI Client

Select your AI client from the available tabs (Claude, Claude Code, Cursor, VS Code, ChatGPT, or Notion). Each tab shows specific setup instructions and a configuration snippet you can copy.

Step 3: Add the Server to Your AI Client

Copy the provided configuration and add it to your AI client. The exact steps vary by client — follow the instructions shown in the setup sheet.

Step 4: Authenticate

After adding the server, your AI client will redirect you to authenticate with your PitchGhost account. During authentication, you'll be asked to:

1. Choose a company — Select which PitchGhost company this AI client should have access to.

2. Set access level — Choose between:

   • Read + Write — The AI client can view and modify your data (create Ghosts, update Finds, trigger scans, etc.)

   • Read Only — The AI client can view your data, but cannot make any changes.

Once approved, your AI assistant is connected and ready to use.

Managing Connections

After connecting, you can manage your MCP connections at any time from Account → Connections → OAuth Integrations.

From this page you can:

• Change the connected company — Switch which company an AI client has access to.

• Toggle access level — Switch between Read Only and Read + Write.

• Revoke access — Disconnect an AI client entirely.

Example Use Cases

Here are a few things you can ask your AI assistant once connected:

Setting up monitoring:

"Create a Twitter Ghost called 'AI Hiring' that looks for tweets from people hiring for AI/ML roles. Set it to scan every 6 hours."

Reviewing opportunities:

"Show me the latest Finds from my LinkedIn Ghost that haven't been reviewed yet. Assign the top 3 to Sarah."

Quick research:

"Search Twitter for recent conversations about enterprise data migration tools and summarize what people are saying."

Managing your pipeline:

"Tag all Finds from this week that mention Series A funding. Mark the ones with confidence scores above 80 as 'Positive' results."

Permissions and Security

• Engage subscription is required to use the MCP integration.

• Admin or Owner permissions are required to delete Ghosts through an AI client.

• Read + Write vs Read Only access is configurable per connection and can be changed at any time.

• All actions performed by the AI assistant respect the same permissions and limits as your PitchGhost account.

• Each AI client connection is isolated — revoking one does not affect others.

• Authentication uses industry-standard OAuth 2.1 with your existing PitchGhost credentials.