Skip to main content

MCP Integration

The Model Context Protocol (MCP) allows AI assistants like Claude and ChatGPT to interact directly with your Promptsy vault. This means you can browse, search, and save prompts without leaving your conversation.

What is MCP?

MCP (Model Context Protocol) is an open standard that enables AI assistants to use external tools and access data sources. When you connect Promptsy to an MCP-compatible AI assistant, you get:

  • Browse your prompts directly in the conversation
  • Save new prompts to your vault without switching apps
  • Search public prompts from the Promptsy community
  • Get full prompt content by ID for detailed viewing
  • Fork public prompts to your personal vault

Available Tools

Promptsy provides five MCP tools:

ToolDescriptionAuth Required
browse_promptsBrowse your personal prompt libraryYes
save_promptSave a new prompt to your vaultYes
search_public_promptsSearch the public prompt communityNo
get_promptRetrieve full content of a prompt by IDYes
fork_promptFork a public prompt to your vaultYes
info

get_prompt and fork_prompt are available in Claude Desktop, Cursor, Windsurf, and other MCP clients. ChatGPT users can access equivalent functionality through the widget interface.

Supported Integrations

PlatformDescriptionGuide
Claude DesktopDesktop app for macOS/Windows/LinuxSetup Guide →
Claude CodeCLI for terminal-based AI codingSetup Guide →
ChatGPT AppsChatGPT web and mobileSetup Guide →
Other MCP ClientsAny MCP-compatible clientGeneric Guide →

Quick Setup

Claude Desktop / Claude Code

  1. Add to your MCP config:
{
"mcpServers": {
"promptsy": {
"url": "https://mcp.promptsy.dev/mcp"
}
}
}
  1. Restart your client
  2. Authorize when prompted

Claude Desktop guide → | Claude Code guide →

ChatGPT Apps

  1. Enable developer mode in Settings → Apps & Connectors → Advanced
  2. Create a new app with URL: https://mcp.promptsy.dev/mcp
  3. Authorize with your Promptsy account

Detailed ChatGPT Apps guide →

How It Works

┌─────────────────┐     ┌─────────────────┐     ┌─────────────────┐
│ AI Assistant │────▶│ MCP Server │────▶│ Promptsy API │
│ (Claude/GPT) │◀────│ mcp.promptsy │◀────│ promptsy.dev │
└─────────────────┘ └─────────────────┘ └─────────────────┘
│ │
│ │
▼ ▼
User sees OAuth 2.1
prompts in Authentication
conversation
  1. You ask your AI assistant to show your prompts
  2. The assistant calls the Promptsy MCP tool
  3. The tool fetches your prompts via the Promptsy API
  4. Results are displayed in your conversation

Authentication

Promptsy uses OAuth 2.1 for secure authentication:

  • First time: You'll be redirected to Promptsy to authorize
  • Scopes: Choose what the AI assistant can access
  • Revoke anytime: Manage connected apps in your Promptsy settings

Learn more about authentication →

Example Usage

Once connected, you can use natural language to interact with your vault:

Browse your prompts:

"Show me my prompts" "What prompts do I have about coding?"

Save a new prompt:

"Save this as a prompt called 'Email Summarizer'" "Add this to my prompt vault"

Search public prompts:

"Find public prompts for writing blog posts" "Search for prompts about data analysis"

Get full prompt content:

"Get the full content of my code review prompt" "Show me the details of prompt abc123..."

Fork public prompts:

"Fork that prompt to my vault" "Save a copy of the email template prompt"

Security

We take security seriously:

  • OAuth 2.1 for secure authentication
  • Prompt injection protection to prevent malicious content
  • Scope-based access - only grant what you need
  • Token expiration - tokens refresh automatically

Learn more about security →

Troubleshooting

Having issues? Check these common problems:

  • Not seeing tools: Restart your AI assistant after configuring
  • Authorization failed: Make sure you're logged into Promptsy
  • Empty results: Check that you have prompts in your vault

Full troubleshooting guide →