google-image-api-skill_skill

Overview

This skill automates extraction of structured image metadata from Google Images using the BrowserAct Google Image API template. It returns clean, machine-ready fields (thumbnails, source, titles, click-through URLs, source logos, product flags) for large-scale visual research. The skill is designed for repeatable, country- and language-biased image harvesting without manual browsing or opening detail pages.

How this skill works

The agent constructs a BrowserAct API request with user-provided keywords, country, language, scroll depth, and item limit. The remote browser session scrolls Google Images, collects list-level entries, and the skill parses each result into standardized fields like thumbnail, link, title, source, source_logo, is_product, related_content_id, and index. It monitors execution logs while running, applies one automatic retry on transient failures, and surfaces explicit guidance when the BrowserAct API key is missing or invalid.

When to use it

• Collect large sets of image thumbnails and metadata without opening detail pages
• Track competitor visual assets and product-style listings across regions
• Build labeled visual datasets for ML or market research
• Monitor localized image trends by country and UI language
• Extract click-through targets for e-commerce discovery

Best practices

• Provide clear, focused search keywords to reduce noisy results
• Set country and language parameters to match the target market for accurate localization
• Limit scroll_count and datelimit to control run time and output size
• Check and set BROWSERACT_API_KEY before invoking; do not run without it
• When a run fails for non-auth reasons, allow one automatic retry, then inspect logs for the exact error

Example use cases

• Daily scrape of SEO keyword image results to analyze visual ranking changes
• Gather competitor product thumbnails and source links for catalog benchmarking
• Compile regional variants of product images by running queries with different country codes
• Extract titles and source logos for a dataset of brand appearance in image results
• Collect structured image metadata to seed a computer vision training dataset