Amazon Search Results Scraper

What it does

Describes a BrowserAct workflow for collecting normalized, source-linked records from Amazon search-result pages that the user is authorized to access.

When to use

Consider it when an authorized browser session is required and the seller accepts the operational, policy, privacy, and stability risks of browser collection.

Inputs

• Authorized Amazon search URL, keyword, category, marketplace, and filters.
• Requested fields, result count, deduplication key, and output format.
• Approved BrowserAct runtime, browser session, proxy, credential, and cost boundaries.

Outputs

• Source-linked search-result records when collection succeeds.
• Normalized ranking, product, price, rating, review, and sponsored fields.
• Explicit access, CAPTCHA, missing-data, or runtime limitations.

Example workflow

1. Define one public Amazon search page and the exact visible fields required.
2. Review BrowserAct permissions and prohibit automatic dependency installation.
3. Run only after approval, then retain source URLs and capture timestamps.

Supported agents and setup

Only the pinned SKILL.md and Skills.sh page were reviewed. The BrowserAct dependency, installer, browser session, and data collection were not executed.

Account, API, and permission requirements

Requires BrowserAct and browser access for live use. Cloud or advanced features may require an account, API key, proxy, or paid usage even though the base Skill does not require them.

Limits and human review boundary

• The repository has no root LICENSE; MIT is stated only in this SKILL.md frontmatter.
• The source instructs automatic BrowserAct installation, which should not run without explicit approval.
• Browser scraping can fail due to access controls, layout changes, CAPTCHA, regional differences, or marketplace terms.
• Cookies, browser identifiers, proxies, credentials, and collected personal data require strict handling.

Pair with Ecommerce Data API

Prefer structured Amazon research and keyword APIs when they cover the task; use browser collection only when authorized visible-page evidence is specifically required.