All notable user-facing changes to Lead Intelligence Enricher. Format follows Keep a Changelog .
For internal architecture decisions and the full development history, see the GitHub repo.
[0.1.9] — 2026-05-28
Added
📊 Three named dataset views in the Apify Console: Overview, Owners found, Sentiment digest. Buyers can scan results without downloading the dataset.
🌍 Multilingual website support: the owner-name extractor now recognizes About / Team / Contact / Impressum pages in German, French, Spanish, Italian, and Portuguese in addition to English.
📑 Open eval methodology: every quality number on the listing is reproducible from the published per-record diffs, labeling guide, and comparator code.
Changed
⚡️ Faster runs: owner extraction and sentiment digest now run in parallel per business. Per-record wall-clock cut roughly in half (~28s → ~15s).
📉 Lower infrastructure cost, same price: model migration drops our cost-of-goods-sold per business by ~5×, with no change to the published $10 per 1,000 pricing. Headroom passed to long-term price stability.
🎯 Two confidence modes: strict (higher precision, more nulls) and balanced (default).
💸 Pay-per-event pricing: $0.005 run start fee + $0.010 per business with ≥ 1 non-null enrichment field. Businesses that produce only nulls are not charged.
🔁 Compass passthrough: every input field (rating, address, phone, hours, etc.) flows to the output unchanged. The actor only adds fields.
🌐 Residential proxy support for website fetching.
💾 Resumable runs: per-business state cached so a re-run skips already-processed records.
Quality methodology
154-business human-labeled eval set spanning US / EU / LATAM / India / Rest
Stratified across categories: restaurants, dentists, plumbers, retail, professional services
Mix includes ~30% no-website, multiple CMS platforms, English + non-English sites
Upcoming
📋 Hand-audit of remaining false-positives to publish the upper-bound precision number.
🔌 Optional output to Apify's request queue for direct downstream-actor wiring.
🌏 Expanded multilingual evaluation set (currently English-heavy).
For bug reports, false-positive examples, or feature requests, drop a note on the actor's Apify Store page . Every reported issue gets added to the next eval iteration.
