BrowserOS: Open Source AI Smart Browser

BrowserOS is an open source AI smart browser developed on Chromium and compatible with all Chrome extensions. It emphasizes privacy, and all data and AI models run locally, with users having the option of using their own API keys or local models (such as Ollama.) BrowserOS offers powerful AI features, including web content analysis, automated tasking, and multi-model comparisons, in an easy to use interface that is consistent with Chrome. It is positioned as Perplexity An open source alternative to Comet for users who value privacy and personalization. Developed by a team in San Francisco under the AGPL-3.0 license, the project is community-driven and encourages user participation and feedback.

Function List

• All Chrome extensions are supported, keeping the familiar browser interface.
• Locally run AI agents with support for Ollama or user-owned API keys to protect data privacy.
• Automate browsing tasks such as auto-filling forms, web navigation, etc.
• Intelligent content analysis, including web page summarization, code parsing and multilingual translation.
• Supports running multiple Large Language Models (LLMs) at the same time and comparing answers from different models with one click.
• Provides visual analysis functions such as screenshot analysis and chart interpretation.
• launch MCP Store that allows one-click installation of automated task modules.
• Plans are underway to launch a built-in AI ad blocking feature to cover a wide range of scenarios.

Using Help

BrowserOS is a desktop app developed on Angular and Electron, combining AI features to create a smart browser experience. Below is a detailed installation and usage guide to help users get started quickly.

Installation process

1. clone warehouse::
   Open a terminal and run the following command to clone the BrowserOS repository:

   git clone https://github.com/browseros-ai/BrowserOS.git
   cd BrowserOS
   

This will download the project files locally and into the project directory.

2. Installation of dependencies ::
   Run it in the project directory:

npm install

This will install Angular, Electron, and other necessary dependencies. Make sure Node.js is installed (the latest LTS version is recommended).
2. Starting the Development Server ::
Use the following command to start the development environment:

npm run electron:serve

This command will also start the Angular Development Server (running in the http://localhost:4200) and Electron applications. Wait a few seconds and the BrowserOS window will open automatically.
Alternatively, it can be started manually:

• In a terminal, run npm start Start the Angular server.
• In another terminal, run npm run electron:dev Start Electron.
  If you need the production environment version, run:

npm run electron:build

2. Verify Installation ::
   Once launched, BrowserOS has a similar interface to Chrome and supports all Chrome extensions. Users can install extensions through the Chrome Web Store or drag and drop the .crx Documentation.

Main Functions

• Using Chrome Extensions ::
  BrowserOS is based on Chromium and supports all Chrome extensions. Users can access chrome://extensions/Click "Load Extracted Extension" to import a local extension, or install it directly from the Chrome Web Store. Once installed, the extension will function just like Chrome.
• Running the Local AI Agent ::
  BrowserOS supports running AI models locally (e.g., Ollama) or connecting to models in the cloud via user-provided API keys.
  • Configuration Ollama ::
    Install Ollama (refer to https://ollama.ai), and then on the BrowserOS settings page, enter the local model address (e.g. http://localhost:11434).
    Once launched, the AI agent can handle tasks such as web content analysis and automated navigation.
  • Using API Keys ::
    Enter the OpenAI, Claude, or other supported model API key on the Settings page. All data processing stays local and browser history is not uploaded to the cloud.
• Automated Browsing Tasks ::
  BrowserOS' AI agent automates web tasks.
  • Open the sidebar and enter a command (e.g. "Fill out this form" or "Navigate to a website").
  • The AI parses the DOM structure of the web page and performs tasks automatically. For example, if you type "Fill out the login form", the AI recognizes the username and password fields and fills in the preset information.
  • Users can customize automation scripts to be saved as MCPs (Modular Task Packages planned for launch).
• Multi-model comparison ::
  BrowserOS allows running multiple large language models (e.g. ChatGPT, Claude, Gemini) at the same time.
  • Select "Multi-model mode" in the sidebar.
  • Enter a question (e.g., "Explain quantum mechanics") and BrowserOS will display answers from multiple models for easy comparison.
  • Supports one-click copying of web page content to the model input box to improve efficiency.
• Intelligent Content Analysis ::
  • Web page summaries : Select web content, right-click and select "Generate Summary", AI will extract key information and generate a short summary.
  • code resolution : Copy the code snippet to the sidebar and the AI can explain the code logic or optimization suggestions.
  • rendering : Select the text, click the "Translate" button, support multi-language real-time translation.
• visual analysis ::
  • Screenshot Analysis: By clicking on the "Screenshot" button in the toolbar, AI will analyze the content of the image, such as identifying graphical data or interpreting problems in the image.
  • Full Page Capture: By selecting "Full Page Analysis", AI generates a visual summary of the entire web page, suitable for complex pages.

Featured Function Operation

• Sidekick LLM ::
  The Sidekick feature of BrowserOS allows quick switching between different models (e.g. ChatGPT, Claude).
  • Click on the model icon in the sidebar to select the desired model.
  • Enter a question or copy the content of a web page and the model will provide contextually relevant answers.
  • Support code discussion, documentation parsing and other technical scenarios.
• Privacy ::
  BrowserOS does not collect user browsing history and all AI processing is done locally. Users can view the open source code (https://github.com/browseros-ai/BrowserOS) to ensure transparency.

caveat

• Ensure that the network is stable to complete the initial dependency installation.
• Local models require high hardware performance, 16GB of RAM and GPU support is recommended.
• Community support: Join the Discord community (link on the GitHub homepage) to get help or suggest features.

application scenario

1. Privacy-sensitive users
   BrowserOS is for privacy-conscious users. All data processing is done locally and browsing history is not uploaded to the cloud. Users can use local models (e.g. Ollama) or their own API keys to avoid the risk of data leakage.
2. Developers & Technology Enthusiasts
   Developers can take advantage of BrowserOS' open source features to customize AI agents or automation scripts. Support for code parsing and documentation analysis is ideal for quickly understanding technical content or debugging code.
3. Efficient information processing
   Researchers or students can use the webpage summary and translation functions to quickly extract the key points of an article or translate foreign language content into Chinese to improve reading efficiency.
4. multimodeling study
   AI researchers can utilize the multi-model comparison feature to test the performance of different models on the same problem, suitable for academic research or model evaluation.

QA

1. Do I need to pay for BrowserOS?
   BrowserOS is completely open source and free to use. Users can choose to bring their own API key or a local model with no subscription fees.
2. Is mobile supported?
   Currently BrowserOS only supports desktop (Windows, Mac, Linux) and is based on Electron. Mobile support is not yet available.
3. How do you ensure privacy and security?
   BrowserOS runs all AI processing locally, code is open source (AGPL-3.0 license), and users can audit the code. Browser data is not uploaded to the cloud.
4. Does it work seamlessly with existing Chrome extensions?
   Yes, BrowserOS is based on Chromium, is compatible with all Chrome extensions, and installs and works the same way as Chrome.