Glassdoor Reviews Scraper & Employer Brand Monitor

Turn public Glassdoor employee reviews into employer-brand and employee-experience research for HR, recruiting and people-analytics teams. Export available ratings, pros, cons and employer replies, then reuse a named watchlist to identify newly observed reviews and company-rating changes.

No Glassdoor login, cookies, or API key is required.

What makes this Actor useful

• Available public review data: overall and six category ratings, headline, pros, cons, advice, recommendation/CEO/outlook signals, job title, current/former status, employment type, tenure, location, date, helpful votes, review URL, and available employer replies. Missing source fields remain null rather than being invented.
• Company rating snapshot: overall rating, total review count, recommend-to-friend, CEO approval, business outlook, and available category ratings.
• Recurring monitoring: named watchlists label every row as baseline, new, or seen and preserve per-review first-seen time and observation count.
• Only-new mode: after the first baseline, emit only newly observed reviews to avoid duplicate dataset charges. Checking the source still uses resources even when there are no new reviews.
• Change signals: identify company-rating fields that changed since the previous complete observation.
• Ready-to-use reports: dataset JSON, formula-safe CSV, Markdown summary, company snapshots, and an honest run report.
• Reliable bounded access: one sessionless specialist Unblocker request per reviews API page replaces the direct, datacenter, and Residential routes that Glassdoor blocked in live tests.
• Safe partial results: blocked or incomplete pages are reported as partial and are never treated as deleted reviews.

Pricing: Saved Reviews + Actor Start

The Actor uses Pay Per Event pricing:

Event	Price
review-scraped	$0.004 per successfully saved review, or $4.00 per 1,000
apify-actor-start	$0.00005 per GB when the Actor starts, minimum one event

Failed company requests and unsaved reviews do not trigger review-scraped. Reviews already saved before a partial failure retain their normal result charges. The Actor Start fee can still apply to empty or failed runs: Apify counts one start event per GB of memory, with at least one event. At the unchanged 256 MB default, that is $0.00005 for the start event.

Platform usage is included under the current Store pricing. The monitoring reports have no separate report event. Check the live Pricing tab for current event prices, discounts and platform-usage terms before running. A saved-review subtotal is not the entire event bill: add the start fee.

Quick start

Use a Glassdoor company Reviews URL:

{
  "companyUrls": [
    "https://www.glassdoor.com/Reviews/Google-Reviews-E9079.htm"
  ],
  "maxReviewsPerCompany": 3,
  "watchlistName": "us-tech-employers",
  "onlyNewReviews": false
}

For recurring monitoring:

1. Run once with onlyNewReviews: false to create a baseline.
2. Save the input as an Apify Task.
3. Schedule the Task with the same watchlistName.
4. Turn on onlyNewReviews when you want later runs to emit only unseen reviews.

Input

Field	Purpose	Default
companyUrls	Up to 20 Glassdoor company Reviews URLs	Google Reviews URL
companyNames	Optional company names to resolve	none
maxReviewsPerCompany	Maximum matching reviews saved per company	3
filterByEmploymentStatus	All, current, or former employees	all
filterByJobTitle	Case-insensitive job-title filter	none
filterByLocation	Case-insensitive location filter	none
watchlistName	Private key separating recurring monitors	default
onlyNewReviews	Emit only reviews unseen in earlier matching runs	false
useUnblocker	Use the recommended bounded anti-blocking route	true
maxRunMinutes	Stop new collection work near this limit	5
proxyConfiguration	Optional custom proxy override	automatic strategy

Direct Glassdoor Reviews URLs are recommended when exact company matching matters.

Review output

Each dataset row can include:

{
  "companyName": "Example Co",
  "overallRating": 4,
  "ratingBand": "positive",
  "ratingWorkLifeBalance": 4,
  "ratingCultureAndValues": 5,
  "ratingDiversityAndInclusion": 4,
  "ratingSeniorLeadership": 3,
  "ratingCareerOpportunities": 4,
  "ratingCompensationAndBenefits": 5,
  "recommendToFriend": "POSITIVE",
  "ceoApproval": "APPROVE",
  "businessOutlook": "POSITIVE",
  "headline": "Supportive remote team",
  "pros": "Great culture and benefits.",
  "cons": "Some workload pressure.",
  "jobTitle": "Software Engineer",
  "employmentStatus": "Current Employee",
  "employmentType": "REGULAR",
  "employmentTenure": "4 years",
  "employmentTenureYears": 4,
  "location": "New York, NY",
  "reviewDate": "2026-10-10",
  "helpfulCount": 17,
  "topics": ["compensation", "culture", "remote-work", "workload"],
  "dataQualityWarnings": [],
  "monitoringStatus": "new",
  "firstObservedAt": "2026-10-11T00:00:00.000Z",
  "observationCount": 1,
  "companyChangeDetected": true,
  "companyChangedFields": ["overallRating", "totalReviewsCount"],
  "sourceQuality": "complete"
}

Topic labels are transparent keyword signals, not AI sentiment scores. ratingBand is deterministic: 4–5 positive, 3–3.9 mixed, below 3 negative.

Saved outputs

• Default dataset: review rows.
• COMPANIES: company rating snapshots.
• MONITORING-REPORT: per-company coverage and change evidence.
• MONITORING-REPORT.md: readable monitoring summary.
• MONITORING-REPORT.csv: spreadsheet-ready review export with formula-injection protection.
• OUTPUT: run status, completed and failed company counts, budget status, and concise failure reasons.

Reliability and cost controls

The Actor uses lightweight HTTP rather than a full browser. The default memory is 256 MB, the automated-test input requests only three reviews, and work is bounded by company, page, response-size, request-time, and total-run limits.

The automatic strategy makes one sessionless Unblocker request per reviews API page. It does not repeat a premium request through an HTML fallback. If you provide proxyConfiguration, that explicit choice replaces the automatic strategy.

Remaining review charges are checked before access setup, company resolution, collection and each dataset push. An exhausted limit is reported as BUDGET_LIMITED; it is not called an access block. A smaller affordable sample is explicitly partial and cannot overwrite a complete monitoring baseline. Startup/report storage costs can still apply.

Readable earlier pages are retained if a later page fails, with partial coverage and no monitoring-baseline update. Repeated pages, failed history writes and unresolved company names are not labeled COMPLETE. CSV and other report exports must succeed before history advances; an error after a history write reports confirmed commits rather than claiming rollback. OUTPUT separates output-billing/report failures from source failures and preserves the count of reviews actually saved.

Each request also respects the remaining total-run time. The run report distinguishes completed companies from unattempted companies. Source access failures remain visible; a total block never becomes a fabricated empty success.

The sessionless Unblocker request window is at most 75 seconds, shortened when the collection deadline is nearer. This allows headroom for slower challenge-solving responses without adding automatic retries, duplicate premium requests or a Residential fallback. Ordinary direct/custom routes retain their existing request limits.

The Unblocker connection accepts the proxy's interception certificate only for that managed route; ordinary direct connections retain normal TLS verification.

Limits

Glassdoor can change its public pages or block access. A run can therefore return partial results or fail for a company. The Actor reports that condition instead of fabricating data or silently claiming completion. Filters may require reading additional public pages before enough matching reviews are found.

This Actor reads public information only. Use the data lawfully, respect applicable privacy and employment laws, and do not use it to make automated decisions about individual employees or job applicants.