What is Cerebellum
Cerebellum is a browser AI intelligent assistant built on Claude 3.5 Sonnet and Selenium WebDriver. It can understand task intent and execute web automation tasks such as data scraping and website automation testing. Cerebellum simplifies web browsing into directed graph navigation, using LLM to analyze page content and interactive elements, intelligently plan action paths, and dynamically adjust strategies based on webpage states and historical operations. Cerebellum supports multiple browsers, can accurately simulate user behavior, and is suitable for complex automation scenarios.
Cerebellum's Main Features
- Graph Navigation: Simplifies web browsing into directed graph navigation, where each webpage is treated as a node and user actions (such as clicks or inputs) serve as edges between nodes.
- Node Discovery: Based on LLM (Large Language Model) analysis of page content, discovers and identifies new webpage nodes.
- Action Decision-Making: Based on the current webpage state and historical user operations, intelligently decides the next action.
- Multi-Browser Support: Compatible with multiple browsers including Chrome, Firefox, Safari, and Edge.
- User Behavior Simulation: Accurately simulates user behavior such as clicks and inputs, suitable for complex automation scenarios.
- Dynamic Strategy Adjustment: Accepts runtime instructions and dynamically adjusts browsing strategies and operations based on real-time feedback.
- Form Filling: Automatically fills web forms with user-provided JSON data.
Cerebellum's Technical Principles
- Webpage Modeling: Models the web browsing process as a directed graph, where each webpage is a node and user actions are connecting edges between nodes.
- LLM Integration: Integrates Large Language Models (LLMs), such as Claude 3.5 Sonnet, to analyze webpage content and interactive elements, identifying and planning new nodes.
- State Analysis: The LLM determines the best next action based on the current webpage state and historical operation data.
- Action Execution and Feedback: Cerebellum executes the operations planned by the LLM and feeds the new webpage state after the operation back to the LLM for the next decision-making step.
- Goal-Oriented: Starting from the initial webpage, the goal is to reach the target node representing task completion.
- Automation Process: Users set goals, and Cerebellum automatically handles the cumbersome script recording and execution, achieving automation tasks.
- Environment Integration: Tightly integrated with Selenium WebDriver, leveraging Selenium's browser driver capabilities to achieve cross-browser automation operations.
Cerebellum's Project Address
- GitHub Repository：https://github.com/theredsix/cerebellum
Cerebellum's Application Scenarios
- Website Automation Testing: Conducts functional testing of webpages and automates the execution of test scripts to ensure that various functions of a website or web application work as expected.
- Data Scraping and Analysis: Automates data scraping from webpages for market research, competitor analysis, or data aggregation services.
- Web Content Management: Manages website content, such as automatically updating articles, publishing blogs, and managing user comments.
- E-Commerce: Automating tasks such as order processing, inventory management, price comparison, and product information updates.
- Customer Service Automation: Improving customer service efficiency based on automated chatbots or FAQs in customer support systems.