# Changelog of Email & Contact Extractor — Emails, Phones, Socials (`haketa/email-extractor`) Actor

- **URL**: https://apify.com/haketa/email-extractor/changelog.md
- **Full Actor documentation**: https://apify.com/haketa/email-extractor.md

## Changelog

### 1.1.0 — 2026-11-01

- **Verification run:** scraped **248 unique company websites** (homepage mode) and a second **74-site deep contact-page run** (6 pages/site) on a broad, real-world list. Zero duplicates on `rootDomain`/`inputUrl`, zero `[object Object]`/malformed-value bugs, core identity fields (`inputUrl`, `rootDomain`, `scrapedAt`) at **100%**, `title` at **80–97%**.
- **Coverage measured:** on the deep contact-page run, **87.8%** of sites returned at least one contact signal — **78.4%** had a social profile and **36.5%** had an email (107 emails harvested across the batch). Social-field coverage on the broad run: LinkedIn 56%, Twitter/X 54%, YouTube 50%, Facebook 49%, Instagram 48%, GitHub 25%.
- **Fix — email noise filter (false positives):** the plausibility check now rejects decoy domains that appear as the **suffix of a subdomain** (e.g. a Sentry DSN key leaking as `<key>@o0.ingest.sentry.io`), and rejects domains with a **hyphen at a label boundary** (the classic `name-replacewith@-domain.com` anti-spam obfuscation some open-source `/team` pages publish). Both rules were validated to reject only noise while keeping every real address, including digit-leading domains like `privacy@1password.com`.
- **README:** refreshed to the gold standard — Output example and field reference regenerated from the live sample, honest per-field coverage notes, expanded FAQ, pagination-and-scale section, and a corrected related-scrapers block.

### 1.0 — 2026-05-31

- Initial release. Email extraction via `mailto:` links, Cloudflare `data-cfemail` decoder, and `name [at] domain [dot] com` / `&#64;` text-obfuscation patterns. Phone numbers via `tel:` links and explicit `+CC` international patterns. Eight social profiles (LinkedIn, Twitter/X, Facebook, Instagram, YouTube, GitHub, TikTok, Pinterest). Multi-page-per-domain crawl across canonical contact paths, concurrent workers, and Apify Proxy with per-request session rotation.