# Changelog of LinkedIn Posts Scraper — Latest Posts (`lergassy/linkedin-posts-scraper`) Actor

- **URL**: https://apify.com/lergassy/linkedin-posts-scraper/changelog.md
- **Full Actor documentation**: https://apify.com/lergassy/linkedin-posts-scraper.md

## Changelog

### 2026-10-13

- "Max cost per run" is now respected: the run writes only as many posts as your cap pays for and then finishes SUCCEEDED with a clear status message, instead of being aborted by the platform. A cap lower than the price of one post opens no page and returns one free notice row.

### 2026-10-09

- Company pages: reposts of other people's posts are no longer lost — the feed is now read card by card (Stripe: 10 posts instead of 7, HubSpot 10 instead of 9, Microsoft 10 instead of 6 on the same page).
- A page still in progress right before the run's timeout is now cut off cleanly and listed in NOT_ATTEMPTED (not charged), so the run ends SUCCEEDED instead of TIMED-OUT.
- Internal: cost-safety switches for the paid routes; run-health module updated.

### 2026-10-05 (2)

- Single post URLs: paste links to posts (`/posts/…` or `/feed/update/urn:li:activity:…`) in the new **Single post URLs** field or the main list — one row per post with every post field.
- New field `topComments`: the top ~8–10 comments LinkedIn shows logged-out visitors (author, text, images, date, likes, link). Always included for single post URLs; new option **Add top comments to posts from pages** adds them to posts found on company and profile pages (one extra page request per post, no extra charge).

### 2026-10-05

- Posts are now read from the page's visible post cards as well, so quote-shares, reposts and posts missing from LinkedIn's search-engine data are returned (Microsoft: 7 posts instead of 0–1; a person's newest item is no longer days behind).
- New fields on every post: comment count, reaction total and top reaction types, images, video, document, link card, hashtags, mentions, links, author headline and photo, quoted post, repost details.
- New option **Include reposts** (on by default).
- A page with fewer posts than asked now adds one free `notice` row explaining how many LinkedIn showed. A company page served without its posts is requested once more.

### 2026-10-03

- Profiles hidden from logged-out visitors now go through an additional unblocking route before the heavier one, so posts from more of them are delivered. Profiles that still do not open remain free.

### 2026-10-02

- Default memory now adapts to your input: short lists use less, long lists get more.

### 2026-10-01

- Profile pages are read through a fresh address for every profile, so long lists are refused far less often.
- Profiles that are still refused get one more try through a heavier unblocking route.
- Company pages are read directly, without the extra route.
- Pages that answer HTTP 999 are no longer retried, so runs finish sooner.
- A run close to its timeout stops cleanly and saves the pages it did not reach in a NOT_ATTEMPTED record; they are not charged.

### 2026-09-29

- Clearer input field descriptions and examples.

### 2026-09-23

- Initial release: latest public posts of any LinkedIn company or person (text, date, likes, post URL, activity id, image, author, follower count), plus articles for people, with no login or cookies.