# Changelog of Weibo Scraper (`sian.agency/weibo-scraper`) Actor

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

## Changelog

All notable changes to **Weibo Scraper** will be documented in this file.

### \[2026-08-28] — Cleaner dataset, clickable links

#### 🧹 Changes

- **One clear Overview** replaces the old Overview / Posts / Users split — every operation now lands in a single curated table, with the complete raw row still one click away via "All fields."
- **Profile, post and video links are clickable** in the dataset instead of plain text.
- **Every row now carries `processedAt` and `scrapingDate`** so you can track freshness without cross-checking the run's start time.

### \[2026-08-25] — Free-tier fair-use limits

#### 🔐 Changes

- **Free plans now run from the Apify Console.** Free-tier runs started through the **API** are declined with an explanatory message and no charge; running from the Console (or an MCP client) is unchanged and still free. Paid plans are unaffected on every entry point.
- **Daily free-run allowance tightened** — free plans get 5 unbilled runs and 8 runs total per day on this Actor, resetting at 00:00 UTC. Ample for evaluating; unlimited free usage now stops. Paid plans are uncapped.
- **Free runs now cap pages, not just rows.** A free run fetches at most 3 result pages, on top of the existing 25-row cap — previously a large `maxPages` on a narrow keyword could keep fetching without ever reaching the row cap. Normal runs are unchanged; paid runs keep the full page range.

#### 🛟 Notes

- Every limit above fails **open**: if the allowance can't be checked, your run proceeds.
- A run stopped by any of these limits finishes **successfully** with a report explaining why — it is not reported as a failure.

### \[2026-08-24]

#### 📊 A run report that shows the data, not just a status line

- **See your first row before you open the dataset** — the run report now previews a real record plus a table of posts or accounts, with pictures and avatars rendered inline, so you can check the field names in one glance.
- **Itemized charges on every run** — a plain statement of what was billed, priced by Apify for your own plan. Failed and empty items appear in their own list, marked $0, each with the one thing to change to get them.
- **One-click retry and one-click integration** — copy the exact input that failed, the exact request for this run, or an MCP config that lets an AI agent run the same job in plain language.
- **Three dataset views in the Console** — Overview, Posts and Users. Pick the layout that matches your operation instead of scrolling a 40-column table.
- **Weibo Scraper is now agent-ready** — connect it through the Apify MCP server and ask for Weibo data conversationally; the README carries a copy-paste prompt for Claude, ChatGPT and Cursor.

#### 💎 User Benefits

- Less time spent working out which of 40+ Weibo fields you actually need.
- Billing you can audit yourself, per run, without opening a support ticket.
- Faster hand-off to a teammate or an AI agent — the report carries the working call, not a screenshot.

#### 🎯 Use Cases

- **Market Intelligence Analyst** checking a China sentiment pull at a glance before exporting it to BigQuery.
- **KOL Agency** sanity-checking a Chinese influencer shortlist from the avatars and follower counts in the Users view.
- **Automation Engineer** lifting the API snippet straight out of the report into an n8n or Make scenario.

### \[2026-07-18]

#### 💰 Predictable spend on free and paid plans alike

- **Free plans get a clear 25-row preview** — free runs return the first 25 rows of the same clean data and then stop, so a trial run can never run up a surprise bill. The run log and the report both say when the cap was reached and how to lift it.
- **Runs stop instead of working unbilled** — if an account cannot be billed, or hits the "Max total charge" you set on the run, the actor stops delivering and tells you which of the two happened.
- **Fewer transient failures** — temporary source hiccups are now absorbed by a longer, jittered retry instead of surfacing as failed rows.

#### 💎 User Benefits

- Test Weibo scraping on a free Apify plan with no billing risk.
- Fewer half-finished runs to re-run and re-pay for.

#### 🎯 Use Cases

- **Journalist / China Researcher** trialling a date-ranged Weibo keyword search before committing budget.
- **Brand Manager** running scheduled Weibo brand monitoring under a fixed per-run charge ceiling.

### \[2026-06-20]

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

### \[2026-05-12]

#### 🎉 Weibo Scraper — Launch!

- **Five Scrapers in One Actor** — extract Weibo post details, user profiles, fan lists, follower lists, and run date-ranged keyword search across Sina Weibo — pick the operation, paste the input, one clean dataset per run.
- **Weibo Post Detail Extraction** — full Weibo post payload by ID: text, posting client, region, topics, @-mentions, repost / comment / attitude (like) counts, all attached pictures, creator info, edit history.
- **User Profile Scraping** — full Weibo user profile by UID: followers, following (friends), statuses (posts), verification reason, gender, location, bio, membership tier, avatar set, cover image — ideal for Chinese KOL discovery.
- **Social Graph Mapping** — paginated fan and follower lists (~20 per page) for any public account. **Unique to this actor on Apify** — no other Weibo scraper exposes the social graph.
- **Date-Ranged Keyword Search** — search posts with optional start/end date (defaults to last 30 days), paginated (~14 posts / page). Supports Chinese, English, and mixed queries.
- **BigInt-Safe 16-Digit Weibo IDs** — Weibo post IDs are preserved at full precision (most scrapers silently corrupt them via `Number.MAX_SAFE_INTEGER` overflow and break cross-operation joins).
- **HTTPS-Normalized Media URLs** — every avatar, cover, picture, and embedded video URL is delivered ready to embed.
- **Graceful Privacy Handling** — accounts that have locked their fan / follower list return a single translated explanatory row, not a crash.
- **Localized Message Translation** — Chinese responses from Weibo (`微博不存在`, `用户不存在`, `博主设置仅针对粉丝展示全部粉丝`) translated to clear English, so a deleted post reads differently from a private follower list.
- **Curated camelCase Aliases + Raw Fields** — clean `weiboId` · `userId` · `screenName` · `repostCount` · `attitudeCount` aliases for everyday users, plus every raw Weibo field spread alongside for power users.
- **HTML Run Report** — every run produces an HTML summary (success rate, errors, duration, inputs) saved to the key-value store, even on fatal crash.
- **Three Predefined Dataset Views** — Overview / Posts / Users — open the right table layout from the Apify Console with one click.

#### 💎 User Benefits

- **The only Weibo actor on Apify with full social graph extraction** — competitors stop at posts and profiles; we expose fan and follower lists for influence network mapping.
- **Date-range search** — track how a topic surged or faded over weeks (rivals search current only).
- **Pay-per-result pricing** — only charged for successful extractions; never billed for empty results or failed requests. Deep single-record operations are priced separately from the high-volume list operations, and the price you pay drops on every higher Apify plan.
- **No account, no API key, no proxies** — paste a UID, post ID, or keyword and run.
- **Cross-operation joins that actually work** — bigint-safe IDs let you pivot from search results → post detail → user profile without losing a single character of precision.
- **GDPR-friendly defaults** — extracts only publicly shared data; respects Weibo's per-user fan / follower privacy settings.

#### 🎯 Use Cases

- **KOL Agency** building a Chinese influencer shortlist for a brand campaign — runs Search Weibo + User Profile across 10 keywords in 30 minutes.
- **Market Intelligence Analyst** at a global FMCG brand monitoring China consumer sentiment before launching a new SKU — schedules daily Search Weibo runs across product-category keywords with date filters.
- **Brand Manager** tracking counterfeit & gray-market mentions on Weibo — hourly Search Weibo with brand keyword + Slack alerts on high-engagement matches.
- **NLP Researcher** building a Chinese-language sentiment dataset — pulls Weibo Detail across hundreds of viral posts in one weekend.
- **Influencer-Marketing Agency** doing due diligence on a Chinese creator — runs Fans + Followers + Profile to map influence network and audience overlap.
- **Crisis Communications Team** monitoring brand crises in China — date-ranged Search Weibo with engagement thresholds.
- **Journalist / China Researcher** tracking topic surges — runs date-window Search Weibo to compare conversation volume across campaigns.