Proxy Lite - An open-source visual language model that supports automated web page tasks.
Proxy Lite is an open-source, lightweight visual language model (VLM) with 3 bytes of parameters, supporting automated web page tasks. Proxy Lite can operate a browser like a human, performing tasks such as web page interaction, data scraping, and form filling.
What is Proxy Lite?
Proxy Lite is an open-source, lightweight Visual Language Model (VLM) with 3 bytes of parameters, supporting automated web page tasks. Proxy Lite can operate a browser like a human, performing repetitive tasks such as web page interaction, data scraping, and form filling, significantly reducing automation costs. Using a three-step decision-making mechanism of "observation-thinking-tool invocation," Proxy Lite boasts strong generalization capabilities and low resource consumption, supporting efficient operation on consumer-grade GPUs. Proxy Lite provides a complete browser interaction framework, supporting various application scenarios such as automated web page testing, data extraction, and intelligent task assistance, making it suitable for developers to quickly deploy and use.
Proxy Lite's main functions
- Web page automationIt can automatically complete various operations in the browser, such as clicking buttons, filling out forms, scrolling pages, and handling pop-ups.
- Web scrapingIt parses web page content, extracts structured data, and supports data scraping from various types of web pages, including news, social media, and e-commerce platforms.
- AI-driven interactive web proxyBy combining visual perception capabilities, we can achieve automated UI testing and front-end behavior analysis.
- Intelligent Task AssistantIt helps users search, filter, and summarize information, improving the webpage navigation experience.
- Low resource consumptionWith only 3B parameters, it supports efficient operation on consumer-grade GPUs without relying on large cloud computing resources.
The technical principles of Proxy Lite
- Visual Language Model (VLM)It combines visual perception (image recognition) and natural language processing capabilities to understand and manipulate web page content.
- Three-step decision-making mechanism:
- observe: Evaluate whether the previous operation was successful and obtain the current status information of the webpage.
- think: Based on the current page state, reason to determine the next action.
- Tool callInteracting with the browser via API, such as clicking, typing, scrolling, etc.
- Browser interaction frameworkIt has a built-in complete browser interaction control framework, uses the Playwright library to drive browser operations, and supports headless and stealth modes to reduce the risk of being blocked by anti-scraping measures.
- Execution feedback mechanismBy drawing on the feedback mechanism of models such as DeepSeek R1, the decision-making process is continuously optimized during task execution to improve execution accuracy.
Proxy Lite project address
- GitHub repository:https://github.com/convergence-ai/proxy-lite
Application scenarios of Proxy Lite
- Web page automationIt automatically completes tasks such as clicking, filling out forms, and scrolling pages, reducing manual operations.
- Web scrapingExtract structured data from news, e-commerce, and other web pages for use in data analysis or content aggregation.
- Automated testingEnables automated UI testing of web applications, quickly detecting interface functionality.
- Intelligent Task AssistantIt helps users search and filter information, improving the efficiency of webpage usage.
- Enterprise-level task automationAutomating internal enterprise processes, such as data entry and cross-system information publishing.