Web Llm MCP Server

🚀 Web-LLM MCP Server

An MCP server that uses Playwright to load and interact with an HTML page containing @mlc-ai/web-llm for local LLM inference.

🚀 Quick Start

To start the MCP server, run the following command:

node index.js

Or you can run the test:

node test.js

✨ Features

• Browser-based LLM: Employs @mlc-ai/web-llm running within a Chromium browser instance.
• Playwright Integration: Automates browser interactions to ensure seamless LLM operations.
• Multiple Tools: Enables text generation, chatting, status checking, model changing, and screenshot taking.
• Model Management: Supports various Web-LLM models with dynamic switching capabilities.

📦 Installation

1. Install the necessary dependencies:

pnpm install

2. Install Playwright browsers:

npx playwright install chromium

💻 Usage Examples

Basic Usage

Start the server:

node index.js

Advanced Usage

Run the test:

node test.js

📚 Documentation

Available Tools

playwright_llm_generate

Generate text using Web-LLM through the browser interface.

Parameters:

• prompt (string): The prompt to generate text from.
• systemPrompt (string, optional): System prompt to set context.
• maxTokens (number, optional): Maximum tokens to generate.
• temperature (number, optional): Temperature for generation (0 - 1).
• model (string, optional): Model to use (will reinitialize if different).

playwright_llm_chat

Initiate an interactive chat session and receive the response.

Parameters:

• message (string): Message to send in the chat.
• clearHistory (boolean, optional): Clear chat history before sending.

playwright_llm_status

Retrieve the current status of the Web-LLM Playwright interface.

playwright_llm_set_model

Alter the current Web-LLM model.

Parameters:

• model (string): Model ID to switch to.

playwright_llm_screenshot

Capture a screenshot of the Web-LLM interface.

Parameters:

• path (string, optional): Path to save the screenshot.

Supported Models

• Llama-3.2-1B-Instruct-q4f32_1-MLC (default)
• Llama-3.2-3B-Instruct-q4f32_1-MLC
• Phi-3.5-mini-instruct-q4f16_1-MLC
• gemma-2-2b-it-q4f32_1-MLC
• Mistral-7B-Instruct-v0.3-q4f16_1-MLC
• Qwen2.5-1.5B-Instruct-q4f32_1-MLC

🔧 Technical Details

The server operates in the following steps:

1. Launches a headless Chromium browser using Playwright.
2. Loads the index.html file, which houses the Web-LLM interface.
3. Waits for the Web-LLM model to initialize.
4. Exposes browser functions via the window.webllmInterface object.
5. Offers MCP tools that invoke these browser functions.

The HTML interface provides a comprehensive Web-LLM implementation with:

• Model initialization and loading progress tracking.
• A chat interface for testing purposes.
• A JavaScript API for programmatic access.
• Error handling and status reporting mechanisms.

💡 Usage Tip

• The first run will be slower as it downloads and initializes the LLM model.
• The browser runs in headless mode by default.
• Screenshots can be taken for interface debugging.
• Model switching requires reinitialization, which takes time.
• The interface is fully self - contained in the HTML file.