MCP SuperAssistant: Bring MCP Tools to ChatGPT, Gemini, and Beyond

MCP SuperAssistant is a Chrome extension designed to integrate Model Context Protocol (MCP) tools directly into the world’s leading AI platforms. Currently, it supports ChatGPT, Perplexity, Google Gemini, Grok, AIStudio, and OpenRouter Chat, with additional platform support currently in development.

Once the extension is installed, you can execute MCP tools within these web interfaces. The extension automatically inserts tool results back into your conversation, creating a more capable and context-aware AI assistant.

What is MCP?

Model Context Protocol (MCP) is an open standard introduced by Anthropic. It acts as a bridge between AI assistants and various data sources, including content repositories, business tools, and development environments. Through this standardized protocol, AI systems can interact with your data securely and in real time.

Key Features

1. Multi-platform support – Compatible with ChatGPT, Perplexity, Google Gemini, and several other major platforms.
2. Integrated Sidebar – A clean, non-intrusive interface that feels like a native part of the AI platform.
3. Tool Detection – Automatically identifies MCP tool calls within AI-generated responses.
4. Single-click Execution – Run detected tools instantly with one click.
5. Result Injection – Smoothly feeds tool outputs back into the chat thread.
6. Auto Mode – Automatically runs detected tools and submits the resulting data to the chat.
7. Content Push Mode – An option to push page content aside rather than using an overlay.
8. Persistent Preferences – Remembers your sidebar positioning, dimensions, and custom settings.
9. Theme Adaptation – Supports both dark and light modes, automatically matching the theme of the AI platform you are using.

How It Works

The process begins when the AI chat interface generates tool-call instructions. MCP SuperAssistant detects these instructions and transmits them to a local MCP proxy server using Server-Sent Events (SSE). The proxy forwards the request to the specific MCP server, which processes the task and returns the results. The data then travels back through the proxy to the extension, which inserts the final output directly into your chat.

The workflow follows this path:

AI Chat Interface → Tool call generation → Extension detects calls → Data sent via SSE → MCP Local Proxy Server → Forwarded to MCP Server → Results returned → Extension receives data → Results inserted into chat

Installation

1. Install from the Chrome Web Store

• Visit the MCP SuperAssistant page on the Chrome Web Store.
• Click the “Add to Chrome” button.
• Confirm the installation prompts.

2. Connect the local proxy server

To enable the Chrome extension to communicate with a local proxy, you can run the MCP SuperAssistant proxy via npx:

npx @srbhptl39/mcp-superassistant-proxy@latest --config ./mcptestconfig.json

Using this proxy is necessary because it facilitates communication with remote MCP servers, provides CORS support for those servers, and offers a health-check endpoint for system monitoring.

Connection steps:

• Launch the proxy server using the command above.
• Open the MCP SuperAssistant sidebar within any supported AI platform.
• Click the server status indicator (which will initially show “Disconnected”).
• Enter your local server URL (the default setting is http://localhost:3006).
• Click “Connect.” The status indicator will switch to “Connected.”

3. Manual Installation (for development)

• Clone the project repository.
• Run pnpm install to set up the necessary dependencies.
• Run pnpm build to compile the extension.
• Navigate to chrome://extensions/ in your browser.
• Enable “Developer mode” in the top right corner.
• Click “Load unpacked” and select the dist directory from your project folder.
• Follow the proxy connection steps mentioned above to link to an MCP server.

Local Development Setup

To get started with local development, you will need Node.js (v16+) and pnpm.

• Install dependencies: pnpm install
• Start development server: pnpm dev
• Build for production: pnpm build
• Generate distribution zip: pnpm zip