Pica MCP Server

A Model Context Protocol (MCP) server that integrates with the Pica API platform, enabling seamless interaction with various third-party services through a standardized interface. Features enhanced intent detection, improved security, and robust code generation capabilities.

Setup Video: https://youtu.be/JJ62NUEkKAs

Demo Video: https://youtu.be/0jeasO20PyM

Features

🔧 Tools

list_user_connections_and_available_connectors - List all available connectors and active connections
get_available_actions - Get available actions for a specific platform
get_action_knowledge - Get detailed information about a specific action including API documentation
execute_action - Execute API actions immediately with full parameter support
generate_action_config_knowledge - Generate secure, production-ready request configurations for code integration

📚 Resources

pica-platform://{platform} - Browse available actions for a platform
pica-connection://{platform}/{key} - View connection details
pica-action://{actionId} - Get detailed action information with knowledge base

Key Capabilities

🎯 Smart Intent Detection

The server automatically detects whether you want to:

Execute actions immediately: "read my emails", "send this message now"
Generate integration code: "write code to send emails", "create a UI for messaging"

🔒 Enhanced Security

Never exposes secrets in generated code
Uses environment variables: PICA_SECRET, PICA_[PLATFORM]_CONNECTION_KEY
Sanitized request configurations for production use

🌐 Multi-Language Support

Generate production-ready code in:

TypeScript/JavaScript
Python
Go, PHP, and more
Auto-detects language from context or asks user

⚡ Production-Ready Code Generation

Real, working HTTP requests (no demo code)
Proper error handling and type definitions
Clean, maintainable code structure
Environment variable best practices

Installation

npm install @picahq/pica-mcp

Deployment Options

Deploy to Vercel

You can deploy this MCP server to Vercel for remote access:

Install dependencies including Vercel adapter:
```
npm install @vercel/mcp-adapter zod
```
Deploy to Vercel:
```
vercel
```
Configure your MCP client to use the remote server:
- For Cursor: https://your-project.vercel.app/api/mcp
- For Claude/Cline: Use npx mcp-remote https://your-project.vercel.app/api/mcp

See DEPLOYMENT.md for detailed Vercel deployment instructions.

Usage

As a Standalone Server

npx @picahq/pica-mcp

In Claude Desktop

To use with Claude Desktop, add the server config:

On MacOS: ~/Library/Application\ Support/Claude/claude_desktop_config.json

On Windows: %APPDATA%/Claude/claude_desktop_config.json

{
  "mcpServers": {
    "pica": {
      "command": "npx",
      "args": ["@picahq/pica-mcp"],
      "env": {
        "PICA_SECRET": "your-pica-secret-key"
      }
    }
  }
}

In Cursor

In the Cursor menu, select "MCP Settings" and update the MCP JSON file to include the following:

{
  "mcpServers": {
    "pica": {
      "command": "npx",
      "args": ["@picahq/pica-mcp"],
      "env": {
        "PICA_SECRET": "your-pica-secret-key"
      }
    }
  }
}

Using Docker

Build the Docker Image:

docker build -t pica-mcp-server .

Run the Docker Container:

docker run -e PICA_SECRET=your_pica_secret_key pica-mcp-server

Installing via Smithery

To install pica for Claude Desktop automatically via Smithery:

npx -y @smithery/cli install @picahq/pica --client claude

Usage Examples

🚀 Execute Actions Immediately

Natural Language:

"Read my last 5 emails from Gmail"

What happens:

Server detects execution intent
Finds Gmail connection
Executes action immediately
Returns actual email data

🔧 Generate Integration Code

Natural Language:

"Create a React UI and write the code to send emails using Gmail"

What happens:

Server detects code generation intent
Gets Gmail send email action knowledge
Generates sanitized request configuration
AI creates production-ready React + API code
Uses environment variables for security

Intent Detection Guide

The server automatically detects your intent based on natural language:

✅ Execute Immediately

"Read my emails"
"Send this message now"
"Get my GitHub repositories"
"Create a new Slack channel"
"Delete this file"

✅ Generate Code

"Write code to read emails"
"Create a UI for messaging"
"Build an app that syncs data"
"Show me how to implement this"
"Generate integration code"

❓ When Unclear

The server will ask: "Would you like me to execute this action now, or generate code for you to use?"

🔐 Required Environment Variables

# Always required
PICA_SECRET=your-pica-secret-key

# Required when generating code for specific platforms
PICA_GMAIL_CONNECTION_KEY=your-gmail-connection-key
PICA_SLACK_CONNECTION_KEY=your-slack-connection-key
# etc.

API Reference

Tools

execute_action

Execute a specific action immediately and return actual results. Use ONLY when the user wants immediate action execution.

When to use: "send this email now", "get my data", "create this item"

Parameters:

actionId (string, required): Action ID
connectionKey (string, required): Connection key
method (string, required): HTTP method
path (string, required): API path
data (object, optional): Request body
pathVariables (object, optional): Path variables for URL templating
queryParams (object, optional): Query parameters
headers (object, optional): Additional headers
isFormData (boolean, optional): Send as multipart/form-data
isFormUrlEncoded (boolean, optional): Send as URL-encoded

Returns:

result: Actual API response data
requestConfig: Sanitized request configuration (no secrets)

generate_action_config_knowledge

Generate secure request configuration for building real integration code. Use when the user wants to build apps, write code, or create integrations.

When to use: "write code", "build an app", "create a UI", "show me how to implement"