Etsy surfaces its best-seller badges and trending tag labels on public product pages, and scraping them at scale is genuinely useful for competitive research, niche validation, and dropshipping product discovery. The catch is that Etsy runs aggressive bot detection, rate-limits unauthenticated crawlers hard, and returns different HTML depending on whether your request looks like a browser or a script. Here is a practical 2026 guide to getting the data reliably.
What Data You Can Actually Pull
Etsy does not expose a public API for best-seller or trending data. Everything you care about lives in rendered HTML or embedded JSON-LD on product and search pages.
Useful fields per listing:
- Listing title, price, sale price
- “Bestseller” badge (a
with classwt-badge--small) - Star rating and review count
- Shop name and sales count
- Tags (visible on listing pages, not search results)
- Estimated monthly sales (inferred from review velocity, not served directly)
Trending tags appear in Etsy’s search autocomplete (/api/v3/ajax/typeahead/etsy/term) and in the “Shop by popular tags” carousels on category pages. Both endpoints are accessible without login but require consistent headers.
How Etsy Detects Bots
Before writing a single line of code, understand the detection stack you are up against:
| Layer | Method | Notes |
|---|---|---|
| TLS fingerprinting | JA3/JA4 hash check | Requests/httpx fail without spoofing |
| Header validation | User-Agent, Accept, Sec-Fetch-* | Missing Sec-Fetch headers = instant block |
| IP reputation | DataDome (embedded on most pages) | Datacenter IPs blocked by default |
| Behavioral analysis | Mouse events, scroll timing | Only triggers on JS-heavy category pages |
| CAPTCHA | hCaptcha | Triggered on rapid listing traversal |
The TLS fingerprint check is the highest-priority hurdle. Plain requests with a spoofed User-Agent still fails because the TLS handshake looks like Python. Use curl_cffi with impersonate="chrome120" or route through a residential proxy with its own TLS termination.
DataDome is the persistent layer. It tracks request cadence across sessions and will silently serve degraded HTML (no badge data, no review count) long before it serves a hard block. This is similar to the detection stack you encounter when doing more general marketplace work like scraping Walmart Marketplace seller data.
Scraping Best-Seller Listings: Working Approach
For listing-level data, the most reliable path in 2026 is:
- Build a seed URL list from Etsy search (
/search?q=)&explicit=1&ship_to=US - Paginate through results pages (up to page 250, ~6000 results per query)
- For each listing URL, fetch the product page and parse the embedded
message me on telegramResources
Proxy Signals Podcast
Operator-level insights on mobile proxies and access infrastructure.Multi-Account Proxies: Setup, Types, Tools & Mistakes (2026)