# Changelog of TikTok Followers and Following Scraper (`fetch_cat/tiktok-followers-scraper`) Actor

- **URL**: https://apify.com/fetch\_cat/tiktok-followers-scraper/changelog.md
- **Full Actor documentation**: https://apify.com/fetch\_cat/tiktok-followers-scraper.md

## Changelog

### 0.1.3 - 2026-08-20

- Fixed default runs failing when an unavailable proxy group was selected internally; enabled proxy runs now use automatic Apify Proxy routing unless you explicitly select an accessible group.
- Batched dataset writes and result accounting to reduce finalization overhead on large relationship exports.
- Added a deadline-aware output checkpoint so a partially saved profile can resume instead of timing out during result persistence.

### 0.1.2 - 2026-07-30

- Added account region when TikTok explicitly exposes it for a public profile.
- Added validated public biography emails and explicitly labelled or linked Instagram, YouTube, and X/Twitter handles.
- Added singular convenience fields, complete arrays, and a dedicated public bio contacts dataset view.

### 0.1.1 - 2026-07-30

- Added full public metadata for related profiles, including signatures, bio links, languages, account timestamps, all avatar sizes, and public profile settings.
- Added friend counts, liked-video counts, organization and commerce metadata, and the complete normalized profile object.
- Added the same rich public metadata for each requested source profile.
- Relationship rows are preserved with `relatedProfileEnrichmentStatus: "base-only"` if optional profile details are temporarily unavailable.

### 0.1.0 - 2026-07-30

- Fixed incorrect generic recommendation results and restored target-specific follower and following exports.
- Added reliable multi-page following and follower pagination for public profiles.
- Added target-total validation, deterministic deduplication, and accurate truncation reporting.
- Improved blocked-profile reliability with bounded fallback attempts that do not extend the overall run deadline.
- Blocked runs now preserve diagnostics, fail clearly instead of reporting a misleading empty success, and do not incur the start charge.
- Improved memory safety for concurrent fallback attempts.
- Added same-run checkpoint recovery for migrations and restarts, including row deduplication and charge-safe continuation.
- Added complete dataset field metadata plus Python, cURL, and MCP usage examples.
- Changed the default first run to 20 relationships per profile for a faster initial result.