Peta MCP Servers

A collection of MCP (Model Context Protocol) servers for Peta Core, created by Dunia Labs

These are MCP servers optimized for efficient and effective use with Peta Core. They can be connected to Peta Core and configured directly for immediate use, or run independently without Peta.

When used with Peta Core, you get:

Secure credential management — API credentials and sensitive information are stored in the vault, never exposed externally, and accessed only when needed
Multi-tenant support — Serve multiple users or organizations from a single deployment
Granular permissions — Assign different access levels and permissions to different API keys
Human-in-the-loop approval — Require explicit human approval for risky or sensitive operations before execution

Available Servers

1. MCP Server Template

Directory: mcp-server-template/

A template for creating new MCP servers. Use this as a starting point for building your own MCP integrations.

Features:

TypeScript-based structure
Testing infrastructure
Best practices and guidelines
Documentation and examples

View Documentation →

2. MCP Google Drive Server

Directory: mcp-google-drive/

An MCP server that provides integration with Google Drive, enabling AI assistants to interact with your Google Drive files and folders.

Features:

File and folder management
Search capabilities
File version history
Docker support
Real-time token updates

View Documentation →

3. MCP PostgreSQL Server

Directory: mcp-postgres/

An MCP server for PostgreSQL database integration, allowing AI assistants to query and interact with PostgreSQL databases safely.

Features:

Read-only and write operations
Schema inspection
Secure query execution
Docker support
Connection pooling

View Documentation →

4. MCP Notion Server

Directory: mcp-notion/

An MCP server for Notion integration, enabling AI assistants to interact with Notion pages, databases, blocks, and users.

Features:

16 tools for Notion operations
Page management (create, read, update, delete, search)
Database operations (query, create, manage)
Block management (append, read, update, delete)
User management
Property management
STDIO transport
Docker support (amd64/arm64)
Token-based authentication

View Documentation →

5. MCP Google Calendar Server

Directory: mcp-google-calendar/

An MCP server for Google Calendar integration, enabling AI assistants to manage calendars and events.

Features:

10 tools for calendar and event management
Calendar management (list, create, delete)
Event operations (create, read, update, delete, search)
Free/busy queries and event creation
STDIO transport
Docker support (amd64/arm64)
Token-based authentication
Runtime token refresh

View Documentation →

6. MCP Figma Server

Directory: mcp-figma/

An MCP server for Figma integration, enabling AI assistants to interact with Figma files, designs, and collaboration features.

Features:

17 tools for Figma operations
File management (list, read, metadata, screenshots)
Design system access (components, styles, variables)
FigJam support (diagram content and screenshots in XML)
Collaboration (comments, replies, version history)
STDIO transport
Docker support (amd64/arm64)
Token-based authentication

View Documentation →

7. MCP REST Gateway

Directory: mcp-rest-gateway/

A configuration-driven gateway that converts any REST API into MCP tools, enabling AI assistants to interact with RESTful services without writing code.

Features:

Convert REST APIs to MCP tools via JSON configuration
Authentication types: bearer, query_param, header, basic, none
Parameter mapping to REST endpoints (path, query, body, header)
JSONPath support for response transformation
Environment variable substitution
OpenAPI generator (auto-generate config from OpenAPI/Swagger)
STDIO transport
Docker support (amd64/arm64)
Compatible with Peta Core

View Documentation →

Repository Structure

peta-mcp-servers/
├── mcp-server-template/     # Template for new MCP servers
├── mcp-google-drive/        # Google Drive integration
├── mcp-postgres/            # PostgreSQL database integration
├── mcp-notion/              # Notion workspace integration
├── mcp-google-calendar/     # Google Calendar integration
├── mcp-figma/               # Figma design integration
├── mcp-rest-gateway/        # REST API to MCP gateway
└── README.md                # This file

Getting Started

Each MCP server has its own documentation and setup instructions. Navigate to the specific server directory and read its README for detailed information.

General Prerequisites

Most servers in this repository require:

Node.js 18+ or compatible runtime
npm or yarn package manager
Appropriate API credentials or access tokens for the service being integrated

Installation

Each server can be installed and run independently:

# Navigate to the specific server directory
cd mcp-google-drive

# Install dependencies
npm install

# Configure environment variables
cp .env.example .env
# Edit .env with your credentials

# Build the server
npm run build

# Run the server (if applicable)
npm start

Docker Images

All MCP servers are available as Docker images on GitHub Container Registry (GHCR):

🐳 Multi-platform support: linux/amd64 and linux/arm64 (Apple Silicon compatible)
📦 View all packages: https://github.com/dunialabs/mcp-servers/packages

Available Images

Server	Image	Latest Version
Google Drive	`ghcr.io/dunialabs/mcp-servers/google-drive`	v1.1.4
PostgreSQL	`ghcr.io/dunialabs/mcp-servers/postgres`	v1.1.2
Notion	`ghcr.io/dunialabs/mcp-servers/notion`	v1.1.1
Figma	`ghcr.io/dunialabs/mcp-servers/figma`	v1.0.1
Google Calendar	`ghcr.io/dunialabs/mcp-servers/google-calendar`	v1.0.1
REST Gateway	`ghcr.io/dunialabs/mcp-servers/rest-gateway`	v1.0.1

Pull Images

# Pull specific version
docker pull ghcr.io/dunialabs/mcp-servers/google-drive:1.1.4

# Pull latest version
docker pull ghcr.io/dunialabs/mcp-servers/google-drive:latest

Using with Claude Desktop

To use these MCP servers with Claude Desktop, add them to your Claude configuration file:

macOS: ~/Library/Application Support/Claude/claude_desktop_config.json

Windows: %APPDATA%\Claude\claude_desktop_config.json

Option 1: Docker via GHCR (Recommended)

Using GitHub Container Registry images provides better reliability and automatic multi-platform support:

{
  "mcpServers": {
    "postgres": {
      "command": "docker",
      "args": [
        "run",
        "--pull=always",
        "-i",
        "--rm",
        "-e",
        "POSTGRES_URL",
        "-e",
        "ACCESS_MODE",
        "ghcr.io/dunialabs/mcp-servers/postgres:latest"
      ],
      "env": {
        "POSTGRES_URL": "postgresql://user:password@localhost:5432/dbname",
        "ACCESS_MODE": "readonly"
      }
    },
    "google-drive": {
      "command": "docker",
      "args": ["run", "--pull=always", "-i", "--rm", "-e", "accessToken", "ghcr.io/dunialabs/mcp-servers/google-drive:latest"],
      "env": {
        "accessToken": "ya29.xxx..."
      }
    },
    "notion": {
      "command": "docker",
      "args": ["run", "--pull=always", "-i", "--rm", "-e", "notionToken", "ghcr.io/dunialabs/mcp-servers/notion:latest"],
      "env": {
        "notionToken": "ntn_xxx..."
      }
    },
    "google-calendar": {
      "command": "docker",
      "args": ["run", "--pull=always", "-i", "--rm", "-e", "accessToken", "ghcr.io/dunialabs/mcp-servers/google-calendar:latest"],
      "env": {
        "accessToken": "ya29.xxx..."
      }
    },
    "figma": {
      "command": "docker",
      "args": ["run", "--pull=always", "-i", "--rm", "-e", "accessToken", "ghcr.io/dunialabs/mcp-servers/figma:latest"],
      "env": {
        "accessToken": "figd_xxx..."
      }
    },
    "rest-gateway": {
      "command": "docker",
      "args": ["run", "--pull=always", "-i", "--rm", "-e", "GATEWAY_CONFIG", "ghcr.io/dunialabs/mcp-servers/rest-gateway:latest"],
      "env": {
        "GATEWAY_CONFIG": "{\"apis\":[{\"name\":\"example-api\",\"baseUrl\":\"https://api.example.com\",\"auth\":{\"type\":\"bearer\",\"token\":\"${API_KEY}\"},\"tools\":[...]}]}"
      }
    }
  }
}

Option 2: Direct Node.js

If you prefer running servers directly without Docker:

{
  "mcpServers": {
    "postgres": {
      "command": "node",
      "args": ["/path/to/peta-mcp-servers/mcp-postgres/dist/index.js"],
      "env": {
        "POSTGRES_URL": "postgresql://user:password@localhost:5432/dbname",
        "ACCESS_MODE": "readonly"
      }
    },
    "google-drive": {
      "command": "node",
      "args": ["/path/to/peta-mcp-servers/mcp-google-drive/dist/stdio.js"],
      "env": {
        "accessToken": "ya29.xxx..."
      }
    },
    "notion": {
      "command": "node",
      "args": ["/path/to/peta-mcp-servers/mcp-notion/dist/stdio.js"],
      "env": {
        "notionToken": "ntn_xxx..."
      }
    },
    "google-calendar": {
      "command": "node",
      "args": ["/path/to/peta-mcp-servers/mcp-google-calendar/dist/stdio.js"],
      "env": {
        "accessToken": "ya29.xxx..."
      }
    },
    "figma": {
      "command": "node",
      "args": ["/path/to/peta-mcp-servers/mcp-figma/dist/stdio.js"],
      "env": {
        "accessToken": "figd_xxx..."
      }
    },
    "rest-gateway": {
      "command": "node",
      "args": ["/path/to/peta-mcp-servers/mcp-rest-gateway/dist/stdio.js"],
      "env": {
        "GATEWAY_CONFIG": "{\"apis\":[{\"name\":\"example-api\",\"baseUrl\":\"https://api.example.com\",\"auth\":{\"type\":\"bearer\",\"token\":\"${API_KEY}\"},\"tools\":[...]}]}"
      }
    }
  }
}

Refer to each server's documentation for specific configuration details.

Creating a New MCP Server

To create a new MCP server in this repository:

Copy the mcp-server-template directory:

cp -r mcp-server-template mcp-your-new-server
cd mcp-your-new-server

Follow the checklist in NEW_PROJECT_CHECKLIST.md
Read TEMPLATE_GUIDE.md for detailed instructions
Implement your server logic in the src/ directory
Update the main README.md to include your new server

Development Guidelines

Each server should be self-contained with its own dependencies
Follow TypeScript best practices
Include comprehensive documentation
Add tests for critical functionality
Use environment variables for configuration
Follow the MCP specification

Contributing

See CONTRIBUTING.md for guidelines on contributing to this project.

Resources

License

This project is licensed under the MIT License - see the LICENSE file for details.

All MCP servers in this repository are licensed under the MIT License.

Support

For issues or questions specific to a server, please refer to that server's documentation. For general questions about this repository, please open an issue.

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

Repository files navigation

Peta MCP Servers

Available Servers

1. MCP Server Template

2. MCP Google Drive Server

3. MCP PostgreSQL Server

4. MCP Notion Server

5. MCP Google Calendar Server

6. MCP Figma Server

7. MCP REST Gateway

Repository Structure

Getting Started

General Prerequisites

Installation

Docker Images

Available Images

Pull Images

Using with Claude Desktop

Option 1: Docker via GHCR (Recommended)

Option 2: Direct Node.js

Creating a New MCP Server

Development Guidelines

Contributing

Resources

License

Support

About

Uh oh!

Releases 1

Packages

Uh oh!

Contributors 2

Uh oh!

Languages

Name		Name	Last commit message	Last commit date
Latest commit History 22 Commits
.github/workflows		.github/workflows
mcp-figma		mcp-figma
mcp-google-calendar		mcp-google-calendar
mcp-google-drive		mcp-google-drive
mcp-notion		mcp-notion
mcp-postgres		mcp-postgres
mcp-rest-gateway		mcp-rest-gateway
mcp-server-template		mcp-server-template
CONTRIBUTING.md		CONTRIBUTING.md
LICENSE		LICENSE
MCP_SDK_1.23.0_UPGRADE_GUIDE.md		MCP_SDK_1.23.0_UPGRADE_GUIDE.md
README.md		README.md
package.json		package.json

License

dunialabs/mcp-servers

Folders and files

Latest commit

History

Repository files navigation

Peta MCP Servers

Available Servers

1. MCP Server Template

2. MCP Google Drive Server

3. MCP PostgreSQL Server

4. MCP Notion Server

5. MCP Google Calendar Server

6. MCP Figma Server

7. MCP REST Gateway

Repository Structure

Getting Started

General Prerequisites

Installation

Docker Images

Available Images

Pull Images

Using with Claude Desktop

Option 1: Docker via GHCR (Recommended)

Option 2: Direct Node.js

Creating a New MCP Server

Development Guidelines

Contributing

Resources

License

Support

About

Topics

Resources

License

Contributing

Uh oh!

Stars

Watchers

Forks

Releases 1

Packages 0

Uh oh!

Contributors 2

Uh oh!

Languages

Packages