Bright Data

**Judgment:** A focused, high-powered scraping and search skill that’s technically strong but carries non-trivial legal/compliance and safety risks due to explicit bot- and CAPTCHA-bypass behavior. **What it does well** - Wraps Bright Data’s Web Unlocker and SERP APIs to: - Scrape arbitrary webpages and return **clean markdown**. - Run **Google searches** and return **structured JSON** (title, link, description). - Offloads proxy rotation, IP management, and anti-bot handling to Bright Data, making it suitable when standard HTTP clients are frequently blocked. - Simple operational setup: configure `BRIGHTDATA_API_KEY` and `BRIGHTDATA_UNLOCKER_ZONE`, then use the provided scripts (`scripts/search.sh`, `scripts/scrape.sh`). **Key risks and constraints** - **Compliance & ToS risk:** Explicitly designed to bypass bot detection and CAPTCHAs. This can violate websites’ terms of service and, in some jurisdictions, raise legal issues; it requires careful policy and legal review before production use. - **Data protection risk:** Can be used to scrape personal or sensitive information at scale if not constrained; there are no built-in guardrails, rate limits, or content filters. - **Third-party dependency:** Reliability and latency depend entirely on Bright Data’s infrastructure, your account configuration, and quota/credit status. - **Operational setup:** Requires correct API key and zone configuration; misconfiguration will lead to tool failures and may be hard for non-technical users to debug. **Recommended scenarios** - You need **high-success-rate scraping** or search where simple HTTP requests or standard scraping libraries are blocked, and you have **clear permission** to access the target sites. - You’re building **LLM tools or RAG systems** that must read live web content or search results and can assume an ops team will manage Bright Data credentials and costs. - You operate in an environment where **legal, compliance, and security teams** have explicitly approved use of Bright Data for scraping. **Avoid or restrict use when** - You lack explicit permission to scrape targets or are subject to strict compliance regimes and have not obtained legal approval. - You need strong, built-in **safety controls** (PII filtering, domain allowlists/denylists, rate limiting) without adding your own enforcement layer. - You prefer first-party, less aggressive fetching (e.g., vanilla HTTP) and can tolerate a higher block rate in exchange for lower compliance risk.