# Changelog of Zhihu Scraper — Q\&A, Answers, Articles, Columns (`sian.agency/zhihu-scraper`) Actor

- **URL**: https://apify.com/sian.agency/zhihu-scraper/changelog.md
- **Full Actor documentation**: https://apify.com/sian.agency/zhihu-scraper.md

## Changelog

All notable changes to **Zhihu Scraper — Q\&A, Answers, Articles & Columns** will be documented in this file.

### \[2026-06-20]

- 🧯 **Graceful gateway-error handling.** Temporary upstream error pages are now retried and routed cleanly instead of surfacing as confusing parsing failures.

### \[2026-05-15]

#### 🎉 Zhihu Scraper — Launch!

- **Keyword Search Across Zhihu** — search answers, questions, articles, and people in one call; mixed-type results dispatched to the right ID/URL schema per row
- **Full Question Answer Threads** — pull every answer for a Zhihu question with **complete HTML body** (not snippets), vote counts, comment counts, and author profile
- **Article Detail with Full HTML** — single-call extraction of any Zhihu article or column post: full HTML content, topic tags, parent column reference, and author credentials
- **Column (Zhuanlan) Article Lists** — paginate the complete catalog of any Zhihu column by slug, ~10 articles per page
- **Author + Badge Data on Every Row** — Zhihu blue/gold badge verification, follower counts, lifetime upvote tallies, headline bios baked into every dataset row — ready for KOL discovery workflows
- **18–19-Digit ID Precision** — IDs preserved as strings so 64-bit Zhihu identifiers never get silently truncated by JavaScript bigint limits
- **HTTPS URL Normalization** — all Zhihu CDN URLs (`*.zhimg.com`) upgraded to HTTPS automatically
- **Resilient Pagination** — built-in retry on transient upstream errors with mandarin-aware error translation (`问题不存在`, `专栏不存在` → plain English)
- **Pay-Per-Result Pricing** — Search $0.004/row, Question Answer $0.005/row, Article Detail $0.040/row, Column Article $0.004/row. Volume discounts auto-applied at SILVER/GOLD/PLATINUM/DIAMOND tiers

#### 💎 User Benefits

- **Gold-standard LLM training data** — full HTML answer bodies plus author credentials and vote signals make this the cleanest source of curated Mandarin Q\&A for SFT and RAG fine-tuning
- **Long-form depth competitors don't ship** — most Zhihu scrapers return excerpts only; we return the entire answer and article body with formatting and images intact
- **One operation, one clean dataset** — no chaining, no manual cursor management, no proxy setup
- **No account, no API key, no Zhihu login** — paste a keyword or ID, click Run, get clean JSON

#### 🎯 Use Cases

- **Wei (ML Engineer, Beijing)** pulls 100K+ Mandarin answer threads per month to build domain-balanced Chinese instruction-tuning datasets
- **Lin (Insights Lead, Shanghai agency)** keyword-tracks branded questions weekly to surface unfiltered consumer sentiment in real Chinese consumer language
- **Anya (PM, B2B SaaS)** monitors competitor mentions in Q\&A threads to catch "X vs. Y" comparison content before it goes viral
- **Marcus (B2B Marketing Lead)** shortlists Zhihu KOLs by topic + badge verification + follower count for sponsored long-form content campaigns
- **Chen (Data Scientist, hedge fund)** runs daily keyword searches across industry watchlists to spot emerging questions for alpha-generation pipelines
- **Dr. Park (Stanford computational linguist)** builds Mandarin discourse corpora and trains discourse-level classifiers on Chinese internet argumentation patterns