How it works

Give your AI new superpowers

Plugins that let your AI do things, not just say things.

You already use an AI assistant. It's great at answering questions and writing text. But what if it could also search your Google Drive, analyze a spreadsheet, deploy your website, or control your smart home? Plugins make that possible. Each plugin teaches your AI a new skill, and you install them with a single click.

You may also see these called MCP servers. MCP (Model Context Protocol) is the open standard that makes plugins work across different AI apps — Claude, ChatGPT, Cursor, and more.

Every plugin on our marketplace is ready to install. No coding required.

What can your AI do?

Creators around the world are building plugins for every use case imaginable.

Productivity

  • Read and draft emails
  • Manage your calendar
  • Search documents and notes

Development

  • Manage code repositories
  • Run deployments
  • Query databases

Creative

  • Generate images
  • Edit and produce media
  • Design assets

Data & Analytics

  • Analyze spreadsheets
  • Build dashboards
  • Run complex queries

Communication

  • Send messages
  • Post updates to channels
  • Manage notifications

And Growing...

  • New tools published every day
  • Built by creators worldwide
  • The ecosystem never stops growing

Two types of MCP servers

Plugins come in two flavors. Both give your AI new abilities, but they run in different places.

Local MCP ServersLocal

Run on your own computer. Your AI app launches the plugin as a small local process whenever you start a conversation. Your data stays on your machine.

  • Data never leaves your machine
  • Installed via npm, pip, or source code
  • Works offline once installed

Remote MCP ServersRemote

Run on a server in the cloud. Your AI app connects to a URL. Nothing to install or update. The creator hosts and maintains the service.

  • Nothing to install: just add a URL
  • Always up-to-date, maintained by the creator
  • Great for cloud services and APIs

Look for the Local and Remote badges on plugin cards. Both types install the same way: paste the config and go.

Get started in 3 steps

No technical knowledge required. If you can install an app, you can use MCP.

1

Choose your AI app

MCP works with Claude, ChatGPT / Codex, Gemini, Copilot, Cursor, and more. If you already use one of these, you are ready to go.

2

Find a tool

Browse our marketplace to discover tools for your workflow. Every tool goes through our adaptive security scan before listing. Filter by category, search by name, or explore what is popular.

3

Click install and follow the steps

We give you the exact configuration to copy and paste, or you can ask your AI assistant to handle the setup for you. Either way it takes about a minute and no coding is required.

How to install an MCP server

It only takes a minute, and you might not even need to do it yourself.

The easy way: let your AI do it

If you use an AI code editor like Cursor, VS Code, Windsurf, or Claude Code, just ask it:

"Add [plugin-name] to my MCP config"

It will find the right file, paste the configuration, and save it for you. No manual steps needed.

Or follow the manual steps below:

1

Find your config file

Each AI app stores its MCP configuration in a specific file. When you click Install on any plugin, we show you the exact file path for your app and operating system. For example, Claude Desktop uses:

Mac

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

Windows

%APPDATA%\Claude\claude_desktop_config.json
2

Open it in any text editor

Open the config file in any text editor. TextEdit (Mac), Notepad (Windows), or VS Code all work. If the file does not exist yet, create a new empty file at that path. If it already has content, you will add to it.

3

Paste the config and save

Copy the configuration from the Install dialog (use the Copy button) and paste it into the file. If the file is empty, paste the entire block. If there is already an mcpServers section, add the new plugin entry inside it. Save the file.

Local plugin (runs on your machine):

claude_desktop_config.json

{
  "mcpServers": {
    "server-name": {
      "command": "npx",
      "args": ["-y", "package-name"]
    }
  }
}

Remote plugin (connects to a URL):

claude_desktop_config.json

{
  "mcpServers": {
    "server-name": {
      "url": "https://example.com/mcp"
    }
  }
}

The Install dialog always shows you the exact config to copy. No guessing needed.

4

Restart your AI app

Close and reopen your AI application. It reads the config file on startup, so a restart picks up the new plugin. You should see the new tools available in your next conversation.

Creator Reputation

Every creator earns a letter grade based on their track record. The grade helps you pick reliable tools from trusted creators.

Grade Scale

A+ / A / A-

Excellent

B+ / B / B-

Good

C+ / C / C-

Fair

D / F

Poor

New creators show as Unrated until they build enough track record.

What Goes Into the Score

Server Uptime

Average uptime across all plugins

User Reviews

Average rating plus bonus for review volume

Response Latency

Average response time

Security Score

Average security score across plugins

Update Frequency

How often plugins are updated

Account Age

Time since account creation

If a signal is unavailable (e.g. no uptime data for local-only plugins), its weight is redistributed to the remaining signals.

Common questions

Do I need to know how to code?

Not at all. Pick a plugin, click Install, and copy the config into your AI app. The whole process takes about a minute.

Is it safe?

Every tool goes through a multi-layer security review before listing. Tools that aren't valid MCP servers — missing protocol implementation, dead repositories, or empty projects — are filtered out automatically. The tools that pass get a public security score so you can make an informed decision before installing.

What AI apps work with MCP?

Claude, ChatGPT, Gemini, Copilot, Cursor, Windsurf, VS Code, and more. Any app that supports the MCP standard can use plugins from the marketplace.

Are plugins free?

Many are free and open source. Premium tools have a price clearly shown on the listing, so there are no surprises.

Why would I pay for a plugin?

Paid plugins tend to be better maintained because creators have a direct incentive to ship bug fixes, new features, and keep up with AI app updates. Free plugins work great too, but paid ones often go further.

What is the difference between local and remote plugins?

Local plugins run on your machine as a small background process, so your data never leaves your computer. Remote plugins run in the cloud and your AI connects via a URL. Either way, setup is the same: paste a config snippet and go.

Who builds these tools?

Independent developers and companies around the world. MCP is an open protocol — anyone can build a plugin and list it here.