Automate browser workflows using vision-based interaction and Playwright.
npx skills add https://github.com/nbiish/native-psionic-programs --skill skyvern-nbiish
Please help me install this Agent Skill. Skill: skyvern Source: https://github.com/nbiish/native-psionic-programs/tree/main/.agents/skills/skyvern Command: npx skills add https://github.com/nbiish/native-psionic-programs --skill skyvern-nbiish
Skyvern provides a self-hosted AI-powered browser automation framework that coordinates a planner, agent, and validator stack with computer vision to interact with pages semantically, replacing brittle manual web interactions.
Run the quickstart wizard to install Skyvern locally and start automating web tasks
High-intent search queries and answers about installing and using this skill.
Browser automation can be achieved using vision-based interaction to navigate pages semantically. This approach uses a Planner, Agent, and Validator architecture with Playwright to execute multi-step web tasks like form filling and data extraction without manual selectors.
Self-hosted AI browser automation supports routing LLM calls entirely through local providers. You can configure local LLM routing using OpenRouter, Ollama, or OpenAI-compatible endpoints to ensure data privacy during web tasks.
Vision-based web automation uses computer vision to interact with page elements semantically rather than through DOM selectors. It coordinates multi-agent workflows to handle complex sequences like login automation, cross-site e-commerce tasks, and data extraction.
Playwright integrates with self-hosted LLMs to enable cross-site e-commerce automation. The framework uses Playwright for browser control while routing AI decisions through local models, allowing multi-agent workflows to operate across different websites.
Run the quickstart wizard to install the framework locally and configure your local LLM provider. This sets up the Planner, Agent, and Validator stack needed to automate login sequences and extract data through vision-based page interaction.
Vision-based browser automation replaces brittle manual selectors but relies heavily on local LLM performance and computer vision accuracy. Complex multi-step workflows may require clear task definitions and optional components like scripts to validate actions accurately.
whmathews15Configure Dex AI model options for cost reduction and offline access.
RightNow-AIAutomates web browsing tasks with Playwright navigation, form filling, and extraction.
funsaizedAutomates deterministic browser tasks for AI agents using ref-based page interactions and multi-session control.
PhucMPhamAutomate browser tasks with Puppeteer scripts for testing and data extraction.
