harpa-grid

What This Skill Does

The harpa-grid skill enables OpenClaw AI agents to interface directly with live web browsers via the HARPA AI Grid REST API. It transforms your local or remote browser instances into programmable endpoints, allowing the agent to perform complex browser automation tasks. The skill supports full-page content scraping, targeted data extraction using CSS and XPath selectors, and executing live web searches. By leveraging the HARPA AI infrastructure, agents can interact with dynamic web content, bypass basic scraping restrictions, and extract structured data from virtually any website while maintaining context for subsequent AI reasoning.

Installation

To integrate this skill, run the following command in your terminal: clawhub install openclaw/skills/skills/alxsharuk/harpa-ai

Ensure you have the HARPA AI Chrome Extension installed, an active browser node configured in the extension's AUTOMATE tab, and your HARPA_API_KEY set as an environment variable in your OpenClaw environment configuration.

Use Cases

• Automated Research: Program the agent to search the web for specific topics, scrape the top search results, and synthesize a summary.
• Price Monitoring: Target specific CSS elements on e-commerce sites to track product pricing across multiple URLs.
• Data Extraction: Scrape structured datasets, contact information, or article metadata from websites for input into downstream data processing pipelines.
• Content Monitoring: Monitor live pages for updates or changes to specific DOM elements.

Example Prompts

1. "Search for the latest documentation on OpenClaw, visit the top 3 results, and summarize the key setup steps into a markdown document."
2. "Go to this URL: [example.com/products] and extract all product titles and prices, then format them into a CSV table."
3. "Check the current stock status for the item on this page using the '.inventory-status' element and tell me if it is available."

Tips & Limitations

• Timeout Management: Always define a reasonable timeout in your API calls to prevent agent hangs on slow-loading pages.
• Selector Accuracy: When scraping, use specific CSS classes or IDs to ensure your data extraction remains robust against layout changes.
• Node Availability: Remember that your agent relies on an active browser node; ensure the HARPA extension is active and that your computer is awake when executing tasks.
• Dynamic Content: HARPA Grid is excellent for JS-heavy sites, making it superior to traditional HTML parsers like BeautifulSoup for modern, interactive web applications.