Glassdoor Company Reviews & Intelligence Scraper

Turns any company's Glassdoor page into a clean, structured Employer Intelligence record — the full firmographic profile, the complete rating breakdown and distribution, and the latest fully-structured employee reviews — from either a Reviews page URL or a company name.

No login. No cookies. No captcha key. No browser. Public data only.

How it works

Glassdoor sits behind Cloudflare's managed Turnstile challenge and behaves like a DataDome-class target that blocks at the IP level. Stealth headless browsers (and even a rebrowser-patched headful Chromium) through residential proxies could not clear it.

This actor instead routes a single plain HTTPS GET through Apify Proxy's UNBLOCKER group, which solves the anti-bot challenge on the proxy side and returns the final, fully-rendered HTML. Glassdoor is a Next.js App Router app, so the data streams inside self.__next_f.push([1,"..."]) chunks (React Server Components payload); the actor concatenates those chunks and extracts the employer profile, rating summary, rating distribution and review objects straight from them. It only ever returns real data — nothing is fabricated.

One cheap request per company

Unlike scrapers that fire 6–12 Unblocker fetches per company chasing sort/filter permutations that Glassdoor simply ignores, this actor does the job in one request — the cheapest and fastest way to a complete snapshot.

Input

Field	Type	Notes
companyUrls	string[]	Glassdoor Reviews page URLs, e.g. https://www.glassdoor.com/Reviews/Google-Reviews-E9079.htm
companyName	string	Plain company name (e.g. Google, Nubank) — searched on Glassdoor
maxReviews	integer	Cap on reviews per company (1–50, default 10) — see the anonymous limit below
includeReviews	boolean	Output individual reviews too (default true); false = overview only
sort	enum	Preferred review sort (best-effort; anonymous Glassdoor usually ignores it)
language	enum	Content language sent as Accept-Language (en, pt, es, fr, de, it, nl)
proxyCountry	string	ISO code for the Unblocker proxy (US, BR, GB, …). Auto-derived from the URL host if empty
proxyGroups	string[]	Apify Proxy groups; defaults to ["UNBLOCKER"]
timeoutSecs	integer	Per-request timeout (20–180, default 90)
debugStoreHtml	boolean	Store the fetched HTML in the key-value store for inspection

At least one of companyUrls / companyName is required.

Output

One company_overview item per company, plus (by default) one review item per review.

company_overview

The most complete company record in the niche:

employer_id, name, legal_name, website, headquarters, industry, sector, size, founded_year, revenue, company_type, description, logo, overall_rating, review_count, rating_count, all_reviews_count, recommend_to_friend_pct, ceo_approval_pct, business_outlook_pct, ratings_by_category (work-life balance, career opportunities, compensation & benefits, culture & values, senior management, diversity & inclusion), and rating_distribution — the full 1★–5★ histogram for the overall score and every category, plus recommend / won't-recommend counts. (Fields absent from a given company's page come back null.)

review

review_id, review_url, title, rating, pros, cons, advice, job_title, location, employment_status, is_current_job, length_of_employment, date, recommend_to_friend (+ raw), ceo_approval (+ raw), business_outlook (+ raw), subratings (6 categories), count_helpful, count_not_helpful, featured, original_language_id, is_translated, and employer_responses — the official company replies (text, date, responder) with has_employer_response.

The anonymous limit — the honest truth

Logged out, Glassdoor's current site serves only the 3 most-recent full reviews per company and hard-blocks every path that would surface more without an account (verified Oct 2026):

• _P2.htm / _P3.htm deep pagination → 502 (refused to anonymous clients)
• ?sort.sortType=… / ?filter.*=… → ignored (identical 3 reviews every time)
• the /graph GraphQL review API → 502 over the Unblocker (needs a logged-in CSRF token)

This actor therefore does not pretend to return hundreds of reviews it cannot legitimately reach without logging in. It positions instead as the richest, cheapest company-intelligence snapshot: one request → a complete employer profile + full rating distribution + the 3 latest structured reviews with employer replies. For bulk historical review mining you would need a logged-in session, which this actor intentionally does not use.