Glassdoor Scraping: Company Reviews, Salaries & Hiring Signals

Glassdoor holds some of the most useful employer data on the web: employee reviews, salary reports, interview experiences, CEO approval ratings, and detailed sub-ratings for culture, work-life balance, and leadership. For anyone doing salary benchmarking, competitive HR analysis, employer-brand research, or investment due diligence, it is a genuinely rich source – and one that has become much harder to access through official means.

The reason is a shift in access policy. Glassdoor, now owned by Recruit Holdings alongside Indeed, restricted its public developer API to enterprise partners in 2024, with no public documentation, no free tier, and no self-serve sign-up (developer reporting, 2026). For most teams the official path is effectively closed. At the same time, the underlying reviews and salary reports remain publicly readable on the site. That combination – valuable public data, no practical API – is precisely why Glassdoor scraping has become a common, if demanding, project in 2026.

This guide covers what Glassdoor data is worth collecting, why it is one of the harder platforms to scrape, the legal considerations specific to employee-generated content, and why human oversight is essential when the data feeds decisions about pay and people.

What Glassdoor Data Contains

Glassdoor's value is in its structure. Each company profile aggregates several distinct data types, most of which are unavailable in comparable depth anywhere else.

Data type	Fields	Primary use
Company reviews	Overall rating, pros, cons, advice to management, review date	Culture research, competitor benchmarking
Sub-ratings	Culture, work-life balance, compensation, leadership, career	Detailed employer comparison
Salary reports	Base pay, additional pay, total comp by role and location	Salary benchmarking, offer calibration
Interview data	Difficulty, process, questions asked, outcome	Hiring intelligence, candidate prep
Company profile	Size, industry, HQ, revenue band, founded year, CEO approval	Firmographics, leadership signals

The salary and sub-rating data is the most sought-after because it is both specific and comparable across companies. A single competitor's compensation bands, or a category-wide view of how employers rate on work-life balance, is the kind of intelligence that directly informs hiring strategy and pay decisions – which is also why the accuracy of the data matters so much.

Why Glassdoor Is Hard to Scrape

Glassdoor sits behind some of the most aggressive defences of any data-rich site, and being realistic about them upfront saves a lot of wasted effort.

Obstacle	What it means	Implication
No public API	Official access restricted to enterprise partners since 2024	Scraping public pages is the practical route
Bot management	A managed bot layer serves challenge pages to naive requests	Simple scripts are blocked immediately
Login wall on salaries	Detailed compensation is gated behind authentication	Full salary access needs authenticated sessions
Behavioural detection	Scroll, mouse, and timing signals flag headless browsers	Sessions must look genuinely human
Structural complexity	Data is embedded in internal JSON / GraphQL responses	Parsing needs care and ongoing maintenance

The login wall on salary data is the pivotal issue. Much of Glassdoor's most valuable information is only fully visible to authenticated users, which moves the project from scraping open pages into the far more sensitive territory of authenticated access. That shift carries real technical, terms-of-service, and ethical weight, and it is the single point where careful human judgement most needs to enter the process.

We cover this specific problem in depth in scraping behind logins: when AI needs human help, and the defences in how anti-bot systems work.

A Practical Approach

Given the obstacles, a sensible Glassdoor project is scoped tightly and executed conservatively rather than run as a broad, aggressive crawl.

1. Scope to what you actually need

Define the exact companies, roles, or categories that matter. Glassdoor punishes broad crawling, and most business questions – how do we compare to five named competitors on pay and culture – only need a focused slice of the data.

2. Prefer public data; treat gated data carefully

Company reviews and overview data are publicly readable and are the safer starting point. Detailed salary data behind the login wall should be approached only with a clear understanding of the terms and ethics involved, and never by harvesting or sharing credentials irresponsibly.

3. Extract the structured payloads

Glassdoor's front end pulls data from internal JSON and GraphQL endpoints. Capturing those structured responses is cleaner than parsing rendered HTML, though it demands maintenance as the site evolves.

4. Normalise and validate

Salary and rating data must be normalised into consistent units and role definitions before it can be compared, and outliers – a single implausible salary report, a review that looks planted – need to be reviewed rather than blindly averaged in.

Normalisation is its own discipline; see data normalization for scraped records.

5. Treat it as a point-in-time snapshot

Employer data ages quickly. Salary bands shift with the market, and a company's rating can swing after a reorganisation, a layoff round, or a leadership change. A Glassdoor pull is a snapshot, not a permanent fact, so benchmarking work should record when the data was collected and refresh it on a sensible cadence rather than citing figures gathered a year ago as if they still hold. Weighting recent reviews more heavily than old ones also gives a truer read of how an employer looks today.

Legal and Ethical Considerations

Glassdoor data raises sharper questions than product or price data because it is employee-generated and can touch on individuals. US courts have generally been more permissive toward scraping publicly available data – the widely cited hiQ Labs v. LinkedIn litigation (resolved 2022) is the reference point – but that does not extend cleanly to login-gated content or to data that could identify individuals. Glassdoor's terms of service also restrict automated access regardless of the legal position.

Ethically, the responsible approach is to stick to aggregate, public, non-identifying data; to avoid attempting to de-anonymise reviewers; and to be especially careful with anything behind authentication. Salary and review data can affect real people's livelihoods, so the bar for care is higher than usual. This article is not legal advice – treat compliance and ethics as project requirements and consult a professional where the stakes justify it.

See is web scraping legal and web scraping ethics for the wider picture.

When Human Review Adds Value

Glassdoor is close to a worst case for pure automation. The defences are strong, the most valuable data sits behind a login wall that raises genuine ethical questions, and the content itself – subjective reviews, self-reported salaries – is noisy and easy to misread. A model left to run unsupervised will either get blocked, cross lines it should not, or produce a benchmarking dataset skewed by fakes and outliers.

An AI-plus-human model addresses each of these. AI handles the collection and structuring of public data at scale; human experts make the judgement calls that this platform demands – deciding what data is appropriate to collect, screening implausible salary reports and planted reviews, normalising roles that do not map cleanly, and interpreting subjective feedback in context. On a dataset that will inform pay and hiring, that human layer is not a luxury.

Try Tendem's AI agent to scope your employer-data task – and escalate to human experts for the judgement calls that Glassdoor's gated, sensitive data requires.

The managed workflow behind this is described on the data scraping service page.

Conclusion

Glassdoor is one of the richest employer-data sources on the internet, and the closing of its public API has made scraping the main practical route to it. But it is also one of the harder and more sensitive platforms to work with: strong anti-bot defences, a login wall around the most valuable data, and content that can touch on individuals. Collecting it responsibly and accurately calls for tight scoping, real care around gated and personal data, and human judgement on both ethics and interpretation.

Delegate your Glassdoor research to Tendem's AI agent – combining automated collection with the expert oversight that salary and review data genuinely need.

Frequently asked questions

Does Glassdoor have a public API in 2026?

No. Glassdoor, now part of Recruit Holdings alongside Indeed, restricted its developer API to enterprise partners in 2024. There is no public documentation, no free tier, and no self-serve developer sign-up. For most teams the official route is effectively closed, which is why scraping the publicly readable pages has become the practical alternative for salary, review, and company research.

Is it legal to scrape Glassdoor?

Scraping publicly visible data generally sits in a more permissive legal area, and the hiQ v. LinkedIn litigation is often cited in support of collecting public data. But that reasoning does not extend cleanly to login-gated content or to data identifying individuals, and Glassdoor's terms restrict automated access regardless. Because reviews and salaries are employee-generated and sensitive, the bar for care is higher. This is not legal advice – consult a professional for anything high-stakes.

Why is Glassdoor so difficult to scrape?

Glassdoor combines a managed bot layer that serves challenge pages to naive requests, behavioural detection that watches scroll and timing patterns, and a login wall around detailed salary data. Its content is also served through internal JSON and GraphQL endpoints that change over time. Together these make simple scripts fail immediately and require careful, well-maintained sessions – and raise real questions about gated data that go beyond the technical.

Can I get Glassdoor salary data without logging in?

Only partially. Company reviews and overview data are publicly readable, but detailed salary breakdowns are largely gated behind authentication. Accessing gated data moves the project into more sensitive territory technically, contractually, and ethically. It should only be approached with a clear understanding of Glassdoor's terms and the ethics involved, and never through irresponsible credential sharing. This is the point where expert human judgement matters most.

What is Glassdoor data used for?

The most common uses are salary benchmarking and offer calibration, competitive HR and employer-brand research, and due diligence – for example, assessing a company's culture and leadership signals before an investment or partnership. Interview data also helps candidates and recruiters. Because the data is subjective and self-reported, it is most reliable when aggregated, screened for fakes and outliers, and interpreted in context rather than taken at face value.