# Changelog of Threads Scraper (`rastriq/threads-scraper`) Actor

- **URL**: https://apify.com/rastriq/threads-scraper/changelog.md
- **Full Actor documentation**: https://apify.com/rastriq/threads-scraper.md

## Changelog

### 0.1 - review round (same version, not yet released)

- Privacy: with `hideAuthorNames`, post and reply `source_url` is `https://www.threads.com/t/<code>` (no username); `ext_profile_id` only on the requested profile row; the biography goes through the email/phone redactor; a private profile outputs only its profile row (`_quality_private_profile = true`). `ext_author_verified` is `null` when the page does not say.
- Levels: `N0+` (curl_cffi chrome) is the base level and `N1` the datacenter proxy; escalation is per target and reversible (a mistyped or removed profile is `unavailable`, it does not pin N1 nor abort the other targets). `levels_used` reports the levels really used.
- Robustness: invalid input or a setup failure writes `SUMMARY` (`invalid_input` / `setup_error`) before failing; a malformed row is dropped and counted (`parse_error`) instead of losing the page; input arrays limited to 50 unique entries and `skipped_inputs` capped at 20.

### 0.1 (2026-11-09)

- First version of Threads Scraper. Reads the public threads.com profile and post pages as an anonymous visitor (Relay JSON embedded in the HTML). No login, no account cookies, no residential proxy, no CAPTCHA, no GraphQL pagination.
- Modes by input: `usernames` (profile row + posts in the public page, about 4 threads) and `postUrls` (post + replies in the public page, about 7). Inputs: `usernames`, `postUrls`, `maxPostsPerProfile`, `maxRepliesPerPost`, `hideAuthorNames`.
- Output: one row per profile, post or reply (`type`), with counters, `published_at` UTC, `author_hash` (HMAC-SHA256 with `AUTHOR_HASH_SALT`), `author_name` (null with `hideAuthorNames`), `ext_*` and `_quality_*` flags (`_quality_truncated_replies`, `_quality_truncated_posts`). Emails and phones redacted; mentions masked with `hideAuthorNames`.
- HTTP by levels: N0 plain httpx, N0+ curl_cffi chrome, N1 same request through an Apify datacenter proxy with a fresh session. One GET per page; session values are never extracted, stored or reused. A page without its JSON or with a login wall counts as `blocked`. Limits: 200 requests, 10 retries, 50 targets per run, 2-3 s between requests, clean stop after 3 failed targets in a row.
- `SUMMARY` record with status `completed | no_results | partial | failed`; PPE event `item-record` with clean stop on the charge limit.
