# Changelog of Welcome to the Jungle Company Scraper — Firmographics (`haketa/wttj-companies-scraper`) Actor

- **URL**: https://apify.com/haketa/wttj-companies-scraper/changelog.md
- **Full Actor documentation**: https://apify.com/haketa/wttj-companies-scraper.md

## Changelog

### 0.2.0 — 2026-11-02

- **Fixed `coverImage` extraction.** WTTJ returns `cover_image` as a localized dict (`{ en: { url }, fr: { url }, … }`); the previous code read `.url` off the wrapper and always produced `null`. It now unwraps the locale (English-first) before reading `.url`, so `coverImage` is populated on ≈99% of companies instead of 0%. Backward-compatible (still handles a direct `.url` or plain-string value).
- **Verified data quality with a fresh run:** 500 unique companies for `searchQuery: "software"` in ~7 seconds, **0 duplicates** on `objectID`. Core-field coverage: `name`/`slug`/`url`/`openJobs`/`objectID` 100%; `size`/`employees`/`sectors`/`offices`/`hqCity`/GPS ≈99%; `logo` 99%. No value bugs (no `[object Object]`, all URLs schemed, numeric fields typed correctly). `tools` present on ~3% of a broad search (companies with a tech page) and `labels` sparse — both data-dependent, not scraper limits.
- **Refreshed the README** to the gold standard: output example and field coverage now match a live run exactly (corrected `slug`, documented `coverImage`), added real match counts and timing, a residential-proxy FAQ, and an extra enrichment use case.
- Confirmed pure-HTTP (impit) crawl against the public Algolia index with datacenter proxy as the cheap default and automatic retry + IP rotation.
