OpenDia: An Open Source Tool to Connect Browsers to AI Models

OpenDia is an open-source project designed to Model Context Protocol (MCPOpenDia supports a variety of features such as page analysis, content extraction, dynamic bookmarking, and dynamic bookmarking. Users can install the OpenDia extension on browsers such as Chrome, Firefox, etc. to automate browser operations in conjunction with a locally running MCP server.OpenDia supports a variety of features such as page analysis, content extraction, dynamic bookmark management, and page style customization. It runs completely locally, with a focus on privacy, and does not need to rely on cloud services. The project is licensed under the MIT license and the code is publicly available on GitHub for developers and technology enthusiasts to explore and contribute. The latest version v1.1 supports Firefox and remote operation, and is more stable.

Function List

• Intelligent Page Analysis: Extracts web page content for processing by AI models.
• Automated operations: support for clicking buttons, filling out forms, and navigating web pages.
• Dynamic bookmark management: Quickly retrieve browsing history based on keywords or topics.
• Page style customization: adjust the appearance of web pages to optimize the user experience.
• Background operation: support multi-tabbed page operation to maintain an efficient workflow.
• Privacy: All operations run locally and no data is uploaded to the cloud.
• Cross-browser support: Compatible with Chrome, Firefox, Edge, Brave and more.
• Remote Control: Supports remote control via different devices or Claude Browser operation.

Using Help

Installation process

To use OpenDia, you need to install both the server side and the browser extension. Below are the detailed steps:

1. Installing Node.js
   OpenDia relies on Node.js (v14 or later). Visit the official Node.js website to download and install the appropriate version for your operating system. Once the installation is complete, run the following command to confirm that the installation was successful:

   node --version
   

2. Cloning of the OpenDia project
   Download the OpenDia source code by running the following command in a terminal or command line:

   git clone https://github.com/aaronjmars/opendia.git
   cd opendia
   

3. Starting the MCP Server
   Go to the server directory and install the dependencies:

   cd opendia-mcp
   npm install
   npm start
   

   By default, the server runs on localhost:5555(WebSocket port). For a customized port, use:

   npx opendia --ws-port=5555 --http-port=5556
   

   If there is a port conflict, OpenDia automatically terminates the existing process and rebinds it.

4. Installing Browser Extensions
   • Chrome
     Open Chrome and go to chrome://extensions/Click on the "Developer Mode" button to enable "Developer Mode". Click "Load unzipped extensions" and select ./opendia-extension/dist/chrome folder. The extension automatically connects to the local server.
   • Firefox Browser
     Open Firefox and go to about:debugging#/runtime/this-firefoxClick on "Load Temporary Add-ons" and select the ./opendia-extension/dist/firefox/manifest.json File. The extension is loaded and ready to use.
   • For other Chromium browsers (e.g. Edge, Brave), see the installation steps for Chrome.
5. Verify Connection
   After the extension loads, confirm that the server is running normally. The extension automatically connects to the localhost:5555If you are using a remote server, configure a tool such as ngrok (see below). If you are using a remote server, configure a tool such as ngrok (see below).
6. Optional: remote operation
   To run on a different device or Claude browser, set up ngrok for port forwarding:
   • Install ngrok (Ubuntu/Debian example):

     curl -s https://ngrok-agent.s3.amazonaws.com/ngrok.asc | sudo tee /etc/apt/trusted.gpg.d/ngrok.asc >/dev/null
     echo "deb https://ngrok-agent.s3.amazonaws.com buster main" | sudo tee /etc/apt/sources.list.d/ngrok.list
     sudo apt update && sudo apt install ngrok
     

   • Get a free authtoken for ngrok (visit the ngrok website to register).
   • Run ngrok:

     ngrok http 5556
     

   • Configure the extended remote connection using the URL provided by ngrok.

Main Functions

• Intelligent Page Analysis
  After installing the extension, open any web page and click the extension icon to enable the "Page Analysis" function, OpenDia will extract the text, images and other contents of the web page for processing by AI models (such as Claude or ChatGPT). Users can send commands via the MCP protocol to have the AI analyze the content and return results. For example, enter "summarize the core idea of this page" and the AI will generate a summary based on the extracted content.
• automatic operation
  OpenDia supports automation through AI commands. For example, if you type "Search for the latest AI trends on Twitter", the extension will automatically open Twitter, enter the search term, and display the results. Users can enter commands in the extension's interface or integrate with local AI models via APIs.
• Dynamic bookmark management
  Select "Bookmarks Management" in the extension interface, enter keywords (e.g. "Machine Learning") or topics, OpenDia will scan the browser history and list related web pages. It supports exporting bookmarks or generating reports, which is convenient for organizing research data.
• Page Style Customization
  Open the extension's Styling Tools and select a preset skin or custom CSS to adjust the fonts, colors, or layout of the web page. This feature is ideal for optimizing the reading experience, such as switching a web page to night mode.
• Running in the background with multiple tabs
  OpenDia supports processing tasks in the background. Users can open multiple tabs at the same time and execute different commands (e.g. search in one tab, fill out a form in another). Running in the background ensures that tasks are not interrupted.

caveat

• Ensure that Node.js and the server are running properly, otherwise the extension will not connect.
• Remote operation requires a stable network to avoid ngrok connection interruptions.
• Extensions require elevated privileges and may trigger browser security warnings, so manually confirm trust.
• The project is open source software, no official technical support, it is recommended to refer to the GitHub community discussion.

application scenario

1. academic research
   Researchers can use OpenDia to quickly retrieve academic articles from their browser history, extract key information, and generate summaries. For example, type "Find machine learning articles read this week" to get a list of relevant web pages and a summary of their content.
2. Automated workflows
   With OpenDia, marketers can automate social media actions such as batch posting content or collecting data. Enter the command "Search LinkedIn for AI-related jobs" to automate searches and organize results.
3. Personalize your browsing experience
   Users can optimize the web page reading experience through the page style customization function, suitable for users who read for a long time or need a high-contrast interface.
4. Development and Testing
   Developers can utilize OpenDia's MCP protocol to develop custom AI tools, test browser automation features, and explore more integration possibilities.

QA

1. Does OpenDia require an internet connection?
   OpenDia core functionality runs locally and does not require an internet connection. Network support is required for remote control or access to online content.
2. What AI models are supported?
   OpenDia supports Claude, ChatGPT and local LLM through the MCP protocol, which requires the user to configure the model.
3. Is it completely free?
   Yes, OpenDia is licensed under the MIT license and is completely free. Users are only responsible for the hardware cost of the runtime environment.
4. How do I fix the problem of the extension not being able to connect to the server?
   Check if the server is running on localhost:5555If the port is not occupied, make sure that the port is not occupied. Restart the server or change the port to solve the problem.