All notable changes to this actor will be documented here.
Initial release. Drop-in compatible with automation-lab/website-contact-finder on input and output schema, at identical pay-per-event pricing ($35/1k actor-start, $1/1k page-crawled).
Quality upgrades over baseline:
- Cloudflare email decoding — decodes
data-cfemail XOR-obfuscated addresses that the baseline silently drops.
- Email obfuscation decoders — handles
name [at] domain [dot] com, (at) / (dot) variants, and HTML-entity-encoded emails (@, .).
- i18n contact-page prioritization — link scoring across English, German, Spanish, French, Italian, Portuguese, Dutch, and Japanese keywords (
impressum, kontakt, contacto, nosotros, contatti, etc.) so the crawl budget is spent on the right pages on non-English sites.
- libphonenumber-js phone parsing — replaces regex with full E.164 validation. Drops false positives from product codes, dates, and SKUs.
- Per-finding source tracking — new
emailSources and phoneSources output fields map each contact to the page URLs where it was discovered. Existing emails / phones arrays unchanged.
- HTML content-type gating —
page-crawled is only charged for successful text/html responses, never for PDFs, images, or failed requests.