# Changelog of Facebook Posts Scraper · No Login · $1.4/1k (`alfalfa/facebook-posts-scraper`) Actor

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

## Changelog

### 2026-09-23 (proxy)

- Runs that use Apify Proxy now always go through the residential proxy group, even when the input selects no group or a different one. Before, an empty group list meant datacenter proxies, which Facebook answers with a login wall or a rate-limit error, so such runs returned nothing. The run log says so in one line.
- No extra cost: you pay per post, and proxy usage is not billed to you.

### 2026-09-18 (post fields)

- Added post format, hashtags, external links, reaction summaries, direct photo/video URLs and counts, a share flag, feed-story ID, page aliases and collection time. Existing fields retain their names, types, values and order.
- Direct media URLs can expire; the Facebook page URL remains in each `media` entry.

### 2026-09-15 (run log)

- Network error messages in the run log no longer include server addresses.

### 2026-09-09

- Facebook rate-limit answers on pagination wait longer before the retry, and every request now gets seven attempts, so a rough proxy minute no longer loses a page.
- Store description and README now name the plain use cases (Facebook posts scraper, n8n, Make, Zapier, AI agents via MCP).
- A page or profile that Facebook does not show without a login (private profile, restricted page, wrong URL) no longer ends the run as failed: the run finishes with a status message saying which URLs could not be scraped, and nothing is charged for them.
- Pages that come back without their posts feed are fetched again before being given up on, so a one-off bad response does not lose the page.
- Video transcripts are retried once when the download is interrupted.
- The scraper always uses a residential proxy on the platform, even when the proxy is switched off in the input (Facebook blocks direct requests, so such runs could only fail).
- Small platform hiccups (input download, billing calls) are retried instead of failing the run.

### 2026-09-08

- Proxy handling hardened: the scraper checks its route at start and switches to a backup route by itself if the primary one fails, so runs finish instead of erroring.
- The proxy country picked in the input now applies to the default residential route as well.

### 2026-09-07

- Output brought fully in line with the official Facebook Posts Scraper, including the link preview, paid partnership flag and live viewer count.
- Reaction names resolved by their stable identifiers, so the breakdown no longer depends on the interface language.
- Private profiles and removed pages fail fast instead of being retried three times.
- Eight ready-made example runs published, from daily monitoring to date-range exports.

### 2026-09-07 (first release)

- Posts from any public page or profile with text, media, reaction breakdown, comments, shares, top comments and co-authors.
- Date filters that stop paging early, so filtered posts are never billed.
- Run budget honoured.