Indeed Company Reviews Scraper - Employee Reviews, Ratings avatar

Indeed Company Reviews Scraper - Employee Reviews, Ratings

Pricing

from $2.00 / 1,000 reviews

Go to Apify Store
Indeed Company Reviews Scraper - Employee Reviews, Ratings

Indeed Company Reviews Scraper - Employee Reviews, Ratings

Scrape Indeed company reviews (employer reviews) by name or URL. One row per employee review: rating, 5 sub-ratings, text, pros, cons, job title, location, employer reply. Free: employer ratings, industry, size. Filters: rating, date, job title, location. Only-new mode for monitoring. 44 countries.

Pricing

from $2.00 / 1,000 reviews

Rating

0.0

(0)

Developer

Tedj MEABIOU

Tedj MEABIOU

Maintained by Community

Actor stats

0

Bookmarked

116

Total users

67

Monthly active users

4 hours ago

Last modified

Features

Limited permissions

Share

Indeed Company Reviews Scraper — every employee review of any company

This Indeed company reviews scraper turns any company page on Indeed into a clean dataset of employee reviews: the 1-5 star rating, the five sub-ratings, the headline and the full text, pros and cons, the reviewer's job title and location, whether they still work there, the date, the helpful votes and the employer's public reply. Give it company names (Home Depot), slugs or /cmp/ URLs and it walks every page for you. It is a plain HTTP scraper — no API key, no login, no browser — so a run of company reviews costs a fraction of what a headless-browser scraper costs, and the same job runs the same way whether you want employer reviews for one company or for a thousand.

Last verified working: 2026-11-19.

Quick start

  1. Type a company name into Indeed companies (Home Depot; no URL needed) and press Start. {"companies": ["Home Depot"]} gave the 100 newest reviews of The Home Depot, the first one 2.7 seconds after the start and all 100 in 7.5 seconds (run 10uif2asbwFQwsnBh).
  2. Download them: Export → view Reviews → CSV or Excel. For a file with one review per line and nothing else, start the run with "statusRows": "none" and "includeCompanyRow": false (see How do I export Indeed reviews to CSV?).
  3. Call it from code or an AI agent with one of these.

API, straight to a CSV of reviews (19 named columns, one review per line):

curl -X POST "https://api.apify.com/v2/acts/kestrel~indeed-company-reviews/run-sync-get-dataset-items?token=<YOUR_APIFY_TOKEN>&format=csv&view=reviews" \
-H 'Content-Type: application/json' \
-d '{"companies":["Google","Basecamp"],"maxReviewsPerCompany":10,"includeCompanyRow":false,"statusRows":"none"}' \
-o indeed-reviews.csv

Run as written on 2026-11-19: 17 lines, every one a review (Google's 10 and all 7 of Basecamp's), 19 columns, 5 seconds from the call to the file (run xnJVWeBf87QbLG8Sn). Add &maxTotalChargeUsd=1 to the URL to cap what a call may cost.

JavaScript (npm i apify-client):

import { ApifyClient } from 'apify-client';
const client = new ApifyClient({ token: process.env.APIFY_TOKEN });
const run = await client.actor('kestrel/indeed-company-reviews').call(
{ companies: ['Starbucks'], sort: 'rating_asc', maxRating: 2, maxReviewsPerCompany: 50 },
{ maxTotalChargeUsd: 0.5 },
);
const { items } = await client.dataset(run.defaultDatasetId).listItems();
for (const r of items.filter((r) => r.type === 'review')) console.log(r.rating, r.job_title, '|', r.title);

Run as written: the 50 lowest-rated Starbucks reviews, all of them one star (run ssgH3BeKHfnHrlNr1). maxTotalChargeUsd caps what the run may cost.

Python: see How do I run the Indeed reviews scraper?.

AI agents (Claude, ChatGPT, Cursor or any MCP client): add Apify's MCP server with this actor as its tool, then ask in plain words, e.g. "What do Starbucks employees complain about? Read the 50 lowest-rated Indeed reviews."

{ "mcpServers": { "apify": { "url": "https://mcp.apify.com?tools=kestrel/indeed-company-reviews" } } }

The hosted server signs you in with OAuth, or takes your Apify token in an Authorization: Bearer header. To run it locally instead (Node.js 22 or later): npx -y @apify/actors-mcp-server --tools kestrel/indeed-company-reviews with APIFY_TOKEN set. The agent gets the tool kestrel--indeed-company-reviews, whose parameters are this actor's input fields; the tool answers with the run's status message and where its rows are, which the agent reads with get-dataset-items. A 5-review call answered in 9.5 seconds (run Gu3xbSFO4ABKoZoLJ), and 100 reviews take about 7 seconds of run time, inside the 30 seconds the server waits for a run. For an agent, "compact": true, "statusRows": "problems" and "includeCompanyRow": false leave only the reviews and anything that went wrong: such a call returned 3 rows of 16 fields (run 7HKZIHZDSSUfpN4rK), where the default also carries a company row and a status row with every company field (run Gu3xbSFO4ABKoZoLJ: 7 rows for 5 reviews).

Indeed 公司评价、员工评论与雇主评分:中文调用与语言设置

按公司名称或 URL 获取员工评论、优缺点、职位、地点和雇主回复。可用于员工体验与雇主研究,不提供求职者资料。

countries=["all"] 查询所有支持地区;国家未指定时应优先使用 all。国家代码按 Indeed 域名,例如 uk、tw、hk;不支持 cn,不会自动把评论翻译为中文。

调用需求参数与结果约定
小规模验证使用下面 JSON;先检查结果和状态,再提高上限
语言与市场按上面的真实字段选择,不要把中文说明当成自动翻译承诺
费用控制maxTotalChargeUsd 是 Apify 运行选项,不是输入 JSON 字段;按定价页的计费事件设置上限
数据完整性检查状态行和 SUMMARY;空结果、来源错误与完整数据集是不同情况
{
"companies": [
"https://www.indeed.com/cmp/Google/reviews"
],
"countries": [
"all"
],
"maxReviewsPerCompany": 1,
"sort": "newest",
"sessions": 1
}

Avis employés / Arbeitgeberbewertungen / opiniones de empleados:按评论实际地区与语言解释。

This localized quick start uses the existing input contract. API users should send the JSON as actor input and set maxTotalChargeUsd separately in Apify run options. Source coverage and translation depend on the platform, not the language of this guide.

Indeed company reviews at a glance

What you give itCompany names (Home Depot, Ernst & Young), slugs (Google) or company URLs (https://www.indeed.com/cmp/Starbucks/reviews), or a keyword and a place to find the employers (trucking in Texas)
What you getOne row per employee review + a free company row + a free status row per company and country
Reviews per page20, walked automatically until the cap or the end
Country domains44, plus all — nearly every review of a company in a single walk
Price$0.003 per review. Company rows, status rows, empty companies and filtered reviews are free
SpeedMeasured on the platform on 2026-11-19 at the default two sessions: first review in about 3 s, 100 reviews in 7 s, 200 from every country in 12 s; 2026-11-17: 300 from three companies in 15 s, 1,000 in 31 s
MonitoringonlyNewReviews on a schedule: each run delivers and bills only what is new since the last one
Needs a browser?No. No API key, no cookie, no login

What does the Indeed Company Reviews Scraper do?

Indeed publishes employee reviews at indeed.com/cmp/<Company>/reviews. Each grades work-life balance, pay and benefits, job security, management and culture separately, has a headline, a body, often pros and cons, and is tagged with a job title, a location and current or former status; employers can reply publicly, and readers vote reviews helpful or not.

This scraper reads all of that, one company at a time, and writes it to a dataset you can export as JSON, CSV, Excel or XML, or read straight from the API. It also collects the aggregate picture Indeed shows in the sidebar — the overall score, the 1-5 star histogram, the five category averages, the Work Happiness score, the topic breakdown and the pros and cons Indeed extracts from the whole corpus — and puts it in one free company row, so you get the company ratings and the individual reviews in the same run.

Typical uses:

  • Employer brand tracking. What current and former staff say about you and your competitors for talent.
  • Employee feedback analysis. Which of the five categories is dragging the score down, from the full text.
  • Recruitment marketing. Quote real strengths; fix the objections candidates read before they apply.
  • Review monitoring. On a schedule with onlyNewReviews, each run delivers and bills only the reviews written since the last one: employer reputation monitoring without a vendor contract.
  • Indeed negative reviews only. Lowest-rated first with a rating ceiling gives a complaints feed you can route to whoever owns the fix.

Indeed scraper for AI agents: which input answers which request?

Use this actor when the request is about what employees say about a company on Indeed: employee reviews, employer ratings, complaints, the five category scores, employer replies, or the company's Indeed profile (industry, size, headquarters, CEO). From 2026-12-02 it also returns the company's salaries by job title, its Q&A (interview questions included) and the business contact details it publishes on its own website, as paid add-ons. Do not use it for Indeed job listings or Glassdoor data — it returns neither.

The user asks forInput
"200 most recent reviews of Home Depot"{"companies": ["Home Depot"], "maxReviewsPerCompany": 200, "countries": ["all"]}
"what do Starbucks employees complain about"{"companies": ["Starbucks"], "sort": "rating_asc", "maxRating": 2, "requireText": true, "maxReviewsPerCompany": 100}
"50 reviews by nurses at Kaiser Permanente in California, last 12 months"{"companies": ["Kaiser Permanente"], "jobTitleContains": "nurse", "locationContains": ", CA", "sinceDate": "12 months", "maxReviewsPerCompany": 50}
"compare Walmart, Target and Costco as employers" (ratings, no reviews){"companies": ["Walmart", "Target", "Costco"], "maxReviewsPerCompany": 1} and read the type: "company" rows
"new reviews of Marriott every morning"{"companies": ["Marriott International"], "onlyNewReviews": true, "countries": ["all"]} on a daily schedule
"what do former Amazon employees say about overtime"{"companies": ["Amazon"], "keywords": ["overtime"], "employmentStatus": "former", "maxReviewsPerCompany": 100}
"Starbucks and its three closest competitors as employers"{"companies": ["Starbucks"], "maxSimilarCompanies": 3, "maxReviewsPerCompany": 200} — the competitors are the ones Indeed lists next to Starbucks
"reviews filed in Malaysia" (no Indeed review site of its own){"companies": ["Google"], "countries": ["my"]}
"what does Google pay software engineers" (from 2026-12-02){"companies": ["Google"], "includeSalaries": true, "maxReviewsPerCompany": 1} and read the type: "salary" rows
"questions people ask about working at Costco" (from 2026-12-02){"companies": ["https://www.indeed.com/cmp/Costco-Wholesale/faq"]} — a /faq URL is its Q&A
"interview questions candidates ask about Google" (from 2026-12-02){"companies": ["Google"], "includeInterviews": true, "maxReviewsPerCompany": 1} and read the type: "question" rows (question_topic: "Interviews")
"email and phone of these employers" (from 2026-12-02){"companies": ["Basecamp", "Acme Widgets"], "includeContacts": true, "maxReviewsPerCompany": 1}
"reviews UK employees left about Tesco, worst first"{"companies": ["Tesco"], "countries": ["uk"], "sort": "rating_asc", "maxReviewsPerCompany": 0}
"employee reviews of trucking companies in Texas" (no company named){"searchCompanies": ["trucking"], "searchLocation": "Texas", "maxCompaniesPerSearch": 5, "maxReviewsPerCompany": 20}: Indeed's own company search picks the employers
"what Amazon employees say about management"{"companies": ["Amazon"], "topic": "management", "maxReviewsPerCompany": 100}: Indeed's own topic filter
"Spanish-language reviews of Google, every country"{"companies": ["Google"], "countries": ["all"], "language": "es"}
"Walmart reviews written in 2025"{"companies": ["Walmart"], "sinceDate": "2025-01-01", "untilDate": "2025-12-31", "maxReviewsPerCompany": 0}
"just the reviews, as a CSV" or "only the reviews, for my context window"{"companies": ["Google"], "maxReviewsPerCompany": 50, "includeCompanyRow": false, "statusRows": "none"}; for an agent, "compact": true and "statusRows": "problems" keep the core fields and only the status rows that report a problem
"reviews for these Indeed URLs" (a list copied from another scraper){"startUrls": [{"url": "https://www.indeed.com/cmp/Google/reviews"}, "https://www.indeed.com/cmp/Starbucks"]}: startUrls and {url} objects are read as companies
"the reviews this Indeed page shows" (a URL with Indeed's own filters){"startUrls": ["https://www.indeed.com/cmp/Tesco/reviews?fcountry=GB&floc=London"]}: the URL's country, city, job title, text search and order are kept
"Amazon reviews worldwide, budget $20"{"companies": ["Amazon"], "countries": ["all"], "maxReviewsPerCompany": 10000} plus Apify run option maxTotalChargeUsd: 20 outside the input. At $0.003 each, that budget allows at most 6,666 delivered reviews ($19.998), even though the result cap is 10,000.

What comes back: one dataset, row types told apart by type. review rows are charged; so are the add-ons' salary and question rows and a company row whose business contacts were found. One free company row per company carries Indeed's own rating, review count, star histogram and category averages. One free status row per company and country says what happened: ok, no_reviews, not_found (with the closest company names Indeed knows in suggestions), duplicate, partial (some reviews delivered, then a page was refused), error, limit (not read: a cap, the spending limit or the time limit came first) or invalid_input (an entry that is not an Indeed company: free, and the rest of your list runs). Every status row also says whether the company is complete, why its walk stopped (stop_reason) and how far back it got (oldest_review_date). A run in which no company resolves ends successfully with the reason on the run and nothing charged. The most a run can cost is companies x maxReviewsPerCompany x $0.003.

How do I run the Indeed reviews scraper?

Type one or more companies into Indeed companies and hit Start; everything else has a working default. The inputs that matter most are companies, countries and maxReviewsPerCompany.

A company can be a plain name, the slug (the part of the URL after /cmp/) or the whole URL, and you can mix them freely. Indeed resolves a name to the page it files the company under — Home Depot becomes The-Home-Depot, Ernst & Young becomes Ey — and every row carries that slug in company:

{
"companies": ["Home Depot", "Google", "https://www.indeed.com/cmp/Starbucks/reviews"],
"countries": ["us"],
"maxReviewsPerCompany": 200,
"sort": "newest",
"includeCompanyRow": true
}

From Python, with the Apify client:

from apify_client import ApifyClient
client = ApifyClient("<YOUR_APIFY_TOKEN>")
run = client.actor("kestrel/indeed-company-reviews").call(run_input={
"companies": ["Google", "Starbucks"],
"countries": ["us"],
"maxReviewsPerCompany": 500,
"sort": "rating_asc",
"maxRating": 2,
"requireText": True,
})
for row in client.dataset(run["defaultDatasetId"]).iterate_items():
if row["type"] == "review":
print(row["rating"], row["job_title"], "|", row["title"])

And from curl, if you just want to download Indeed reviews as CSV in one command. view=reviews gives the 19 review columns a spreadsheet wants (without it the CSV also flattens the company row's nested fields into every line: 987 columns, run X6oDH8IMJdfAilihc), and "statusRows": "none" with "includeCompanyRow": false keeps the company and status rows out, so every line is a review:

curl -X POST "https://api.apify.com/v2/acts/kestrel~indeed-company-reviews/run-sync-get-dataset-items?token=<YOUR_APIFY_TOKEN>&format=csv&view=reviews" \
-H 'Content-Type: application/json' \
-d '{"companies":["Google"],"countries":["us"],"maxReviewsPerCompany":100,"includeCompanyRow":false,"statusRows":"none"}' \
-o indeed-reviews.csv

Each line is one review (test test_status_rows_none_keeps_the_dataset_to_reviews_and_the_status_record_lists_every_company); the run's STATUS record and SUMMARY still say what happened to each company. Without statusRows the last line of each company is its status row (no review_id), as in run e2lhXxxI8PmxFLkh5 (19 columns, 100 reviews and the status row).

Input reference

Every field below is optional except companies (or one of the target lists other scrapers use, see Inputs from other scrapers work too), and every one of them has a sensible default.

Which companies

  • companies — company names, Indeed slugs or /cmp/ URLs. A name is resolved by Indeed itself and the rows carry the slug it chose; a name Indeed does not know costs nothing and comes back as not_found with the closest names it does know. A URL that names a country domain (https://ca.indeed.com/cmp/Google) is read on that domain only; anything else uses the countries list. Two spellings of one company (Home Depot, The-Home-Depot) are recognised and billed once. A company's own website works too (https://www.lacoste.com, www.apple.com/en-us): it is read under the Indeed company whose own Indeed page lists that site — the first one in Indeed's order, never a near miss — and the status row's note says so; when no company Indeed suggests lists it, the entry comes back not_found with those companies and their websites in suggestions, free. A bare domain (lacoste.com) is tried as a name first, as Indeed resolves some (Amazon.com), then by its website. An entry that is not an Indeed company (a Glassdoor link, an empty line) gets a free invalid_input status row and the rest of the list runs; only a list with no Indeed company at all ends as an input problem.
  • searchCompanies (+ searchLocation, maxCompaniesPerSearch) — find the employers with Indeed's own company search, the one behind "Popular trucking companies in Texas" on indeed.com: one keyword per search (an industry, a trade, a kind of employer) and an optional place. Each search reads the first maxCompaniesPerSearch companies that have reviews (default 10, Indeed lists at most 50), in Indeed's order and exactly like companies you name: the same filters, caps and countries. The search's own free status row (target: "search: trucking in Texas") lists every company Indeed showed, with its rating and review count, in search_results, and each company row says which search found it in found_by. A company you also named is read once, and a listing named exactly like the keyword with fewer than 100 reviews, which Indeed puts first ("Trucking", 3 reviews, for trucking), is skipped as a namesake. Searching is free; a search that lists nothing is a free not_found row. Indeed's country sites search too (uk, de, ca and in answered 50 each on 2026-11-19): the search runs on the site of the first of your countries.
  • countries — which Indeed domains to read each company on. Default ["us"]. See the section below. When the countries you chose hold at most half of a company's reviews, the run card, SUMMARY outside_countries and the company's status row say how many were left out.
  • A pasted review URL keeps Indeed's own filters: ?fcountry=GB reads the UK reviews whatever countries says (as a uk.indeed.com URL does), and floc (a city), fjobtitle (a job title, Indeed's normalised one: "Software Engineer" includes the reviewer who wrote "SDE"), ftext (Indeed's own text search), ftopic, fjobcat, lang and sort go with every page, so you get the reviews that page shows. Indeed gives the filtered count for job-title, text and city views (found_review_count: 242 Google software engineers in the US), but counts a lang view as the whole company, so such a row leaves found_review_count empty and says why in note. Your own filters apply on top of them; your sort, when you set one, wins over the URL's. Two URLs of one company with different filters are two targets, and a review they share is billed once.

Which reviews

  • maxReviewsPerCompany — 0 takes every review the chosen domains list; N delivers the first N reviews that pass your filters, in the chosen order, counted across all countries: 50 with jobTitleContains: "nurse" is 50 nurse reviews, not the nurses among the first 50. Default 100. This is the main cost control.
  • sort — newest (default), helpfulness, rating_desc (highest first) or rating_asc (lowest first). Common spellings from other tools are read too (lowest, highest, date, most_recent, helpful, relevance = newest); anything else ends the run before a request is made. Indeed itself silently ignores an unknown sort and serves newest first, which is why the actor checks it.
  • minRating / maxRating — keep only reviews at or above / at or below a 1-5 star rating. 0 means no bound.
  • requireText — drop rating-only reviews.
  • jobTitleContains — case-insensitive substring of the reviewer's job title, e.g. engineer, nurse, driver.
  • locationContains — case-insensitive substring of the review's location, e.g. Mountain View, TX.
  • keywords — keep a review whose title, text, pros or cons contains any of these words or phrases, e.g. ["overtime", "management"].
  • employmentStatus — current (still works there), former (left) or any (default).
  • sinceDate — YYYY-MM-DD or a relative string such as 30 days, 2 weeks, 12 months, so a saved input never goes stale.
  • onlyNewReviews — monitoring mode. The actor remembers what it delivered (per company, country and filter set) and each later run delivers, and bills, only reviews it has not given you before. The first run is the baseline: the newest maxReviewsPerCompany reviews. Needs the default newest order.
  • stateKey — with onlyNewReviews: a name for this monitor's memory ("hr-team"), so two monitors of the same company and filters keep separate memories and each gets every new review. A new name starts a new baseline. Empty = one memory per company, country and filter set, as before.
  • topic — only the reviews Indeed files under one topic, with Indeed's own filter: work_life_balance, compensation_benefits, job_security_advancement, management or culture_values (other tools' spellings such as workLifeBalance, payAndBenefits or culture work too). Google, US, management: 994 of its 4,068 US reviews (2026-11-19). Indeed applies it, so only those pages are read, and the status row's found_review_count is the topic's count.
  • language — only reviews written in one language, with Indeed's own filter: en, es, de, fr … (or Spanish). With countries: ["all"] it collects a language worldwide. Indeed counts no reviews per language, so found_review_count stays empty and the note says why.
  • untilDate — only reviews published on or before a date (2025-12-31, or 6 months): with sinceDate, a date window. The newer reviews are read and dropped before billing, never charged. Not with onlyNewReviews.
  • includeCompanyRow — emit the free company row (default on). This costs one extra free request per company, to the company overview page, which is where industry, size, headquarters, website, founded year, revenue band, CEO, CEO approval, the yearly ratings, the happiness scores, open jobs and similar companies live.
  • maxSimilarCompanies — also read the first N companies Indeed lists as similar to each company you give (default 0 = off; its 1.7 name includeSimilarCompanies, a number or true = 3, works too). They are read like your own companies, under the same cap and filters; their status rows say similar to <company> in note, and their company rows carry similar_to.
  • includeSalaries / maxSalariesPerCompany — add-on, from 2026-12-02: the company's salary rows, one per job title with a reported salary (default off; cap 100 per company, 0 = all).
  • includeQuestions / maxQuestionsPerCompany — add-on, from 2026-12-02: the company's Q&A rows (default off; cap 50 per company, 0 = all).
  • includeInterviews — add-on, from 2026-12-02: the interview questions people asked about the company, each with its top answer (the Interviews topic of its Q&A), as Q&A rows with question_topic: "Interviews" (default off). It shares maxQuestionsPerCompany with includeQuestions; with both on, the interview questions come first and no question is delivered twice.
  • includeContacts — add-on, from 2026-12-02: the business contact details the company publishes on its own website, on its company row (default off).
  • notifyWebhookUrl (+ notifyWebhookHeaders, notifyLimit, notifyAlways) — when the run ends, post what it delivered to your own Slack, Discord or Telegram, or as JSON to n8n, Make or Zapier: the review count per company and the first reviews with rating, title and link. With onlyNewReviews on a schedule that is a new-review alert, silent on a quiet day. A webhook that wants a token gets it from notifyWebhookHeaders ({"Authorization": "Bearer …"}, or one Name: value header). Free; the URL and the headers are stored as secret inputs.
  • resumeFromRun — the id of a run that stopped before every company was read: at its time limit, at its spending limit (maximum cost), after a graceful abort, or where Indeed refused a company. The same input with this field continues each company where that run stopped, under the same per-company cap, and delivers nothing twice. Such a run says so on its run card and names its own id for this. A run aborted at once (no grace period) saves nothing there: resurrect it instead (see the FAQ).
  • failOnIncomplete — off by default. On: the run ends FAILED when a company was cut short by something the input did not ask for (the spending limit, the time limit, Indeed refusing it, a name not found, an entry that is not an Indeed company), so an automation cannot take a short dataset for a complete one. Your own caps and date limits never count, and what was delivered stays delivered and billed.

Output

  • compact — review and company rows carry only their core fields (a review: review_id, company, rating, title, text, pros, cons, job_title, location, employment_status, review_date, url; a company: rating, review count, industry, size, headquarters, website, CEO and any contacts), for LLM context windows and MCP tools. Status rows and add-on rows stay whole; the price is the same.
  • excludeEmptyFields — leave out null, empty-text and empty-list fields from every row (CSV columns then vary from row to row).
  • statusRows — which free status rows go into the dataset: all (default, one per company and country), problems (only companies not found, without reviews, cut short, refused, given twice or missing most of their reviews in the chosen countries) or none. With problems or none every status row is in the run's STATUS record (key-value store) and SUMMARY counts them all, so nothing is lost; none plus includeCompanyRow: false makes a dataset of reviews only, one per line in any export.

Performance

  • sessions — how many proxy sessions (egress IPs) run in parallel, each carrying up to three downloads at once. Default 2.
  • perIp — requests per second per session. Default 1.
  • proxyConfiguration — keep the default. The actor tests your account's datacenter IPs, keeps the ones Indeed serves (and remembers them for your next run), switches to residential by itself only when none of them is served, and moves back when one is.

Two notes on where the filters run. minRating, maxRating, requireText, jobTitleContains, locationContains and sinceDate are applied inside the actor, before billing, so a review a filter drops is never charged. sort is applied by Indeed itself. The combination is what makes a complaints feed cheap: sort: "rating_asc" with maxRating: 2 puts the one- and two-star reviews on the first page and stops the walk as soon as a whole page is above the ceiling.

Inputs from other scrapers work too

Inputs written for other Indeed review scrapers, or by AI agents used to them, run as they are. Company lists are merged: companies, startUrls (strings or {"url": …} objects, the Apify start-URL shape), companyNames, urls and url all add targets. Other settings are read under our name only when our own field is left unset (absent, empty or at its default); otherwise ours wins and theirs is listed as ignored.

Their fieldRead as
maxReviews, maxNumberOfReviewsmaxReviewsPerCompany (per company, 0 = all)
targetDate, since, reviewsSincesinceDate; maxAgeDays: 30 = sinceDate: "30 days"
monitoringModeForReviews, incrementalMode, newReviewsOnlyonlyNewReviews
reviewsSort, sortBy (newest, highest, lowest, relevance, …)sort
country, reviewsFromCountry (us, uk, GB, ALL, www, All countries, worldwide, uk.indeed.com, USA, United Kingdom, …)countries; scrapifier's reviewsFromCountry (the reviewers' country) is read before its country (the Indeed site, www = the US one)
includeInterviews, extraData: ["interviews"], sections: ["interviews"]includeInterviews (interview questions, from 2026-12-02)
location, jobTitlelocationContains, jobTitleContains
includeOverviewincludeCompanyRow
includeReviews: falsecompany rows only: page one of each company is read, no review is delivered or charged
maxResultsmaxItems (a total for the run)
webhookUrl, slackWebhookUrl, discordWebhookUrl, telegramToken + telegramChatId, notificationLimitnotifyWebhookUrl, notifyLimit (blackfalcondata's alerts)
maxItems with memo23's own fields (monitoringModeForReviews, targetDate, maxConcurrency, enrichEmails, …)memo23's meaning, per start URL: maxReviewsPerCompany
CompanyUrls, reviewsUrl, companyKeys, searchQueriescompanies (getdataforme, shareze001, kaix and aboo.script)
itemLimitmaxItems (a total for the run)
max_page, sizemaxReviewsPerCompany (max_page is pages of 20; 0 = all)
dateFrom, untilsinceDate, untilDate
reviewCountry, market, domain, allCountries, all_countriescountries (all_countries: true = ["all"] and wins over the same form's domain; reviewCountry: "market" reads market)
topic, language, stateKey, compact, excludeEmptyFieldsthe same names here, with every spelling of the topic other tools use
includeCompanyInfo, qincludeCompanyRow, keywords
deliveryWebhookUrl, webhookHeadersnotifyWebhookUrl, notifyWebhookHeaders
maxSecondsa time limit of its own: the run stops starting pages a few seconds before it, keeps and bills what it read, and saves a resumeFromRun point
kaix's keyword (+ location, maxItems) with no company givena company search (searchCompanies, searchLocation, maxCompaniesPerSearch) giving company rows only, free, as kaix sells company profiles; set maxReviewsPerCompany to get reviews as well. kaix's companyKeys likewise give company rows only
includeEmpty, emitInputStatus, flatten, notifyOnlyChanges; memo23's includeMoreJobDetails and includeReviewStatshow this actor always works (a status row for every company, SUMMARY, flat review rows, alerts of what was delivered; from 1.18 each job row reads the job's own page, and the free company row carries Indeed's review stats): listed in aliased_fields, never as unknown

A URL of a company's sub-page asks for that part of the company: /salaries for its salary rows, /faq for its Q&A rows and /interviews for its interview questions (the add-ons, from 2026-12-02 — until then, and for /jobs and /about, the company's reviews are read and its status row says so in note). A review URL's own filters (fcountry, floc, fjobtitle, ftext, sort …) are kept. Several sub-pages of one company are one company, read once. silentflow's extraData: ["salaries", "faq"], tidyharvest's sections and includeContacts, and memo23's enrichEmails switch the add-ons on too. Fields this actor does not know or offer (proxy, maxConcurrency, blackfalcondata's emitUnchanged and emitExpired, MCP-connector and WhatsApp fields, …) are dropped and listed in SUMMARY ignored_fields and at the end of the run card, unless their form sent them empty or switched off; what was read under another name is listed in aliased_fields.

Is this the best Indeed company reviews scraper? The audit of 2026-11-19

Two tables. The first takes what users of the other Indeed and Glassdoor review scrapers asked for, praised or complained about (37 text reviews and 57 issue threads of 20 actors, read 2026-11-18 and re-read 2026-11-19) and measures this actor on each. The second, in the next section, takes every feature of the five Indeed company scrapers with the most users. Verdicts: ✓ better than every rival measured, = as good as the best of them, ✗ behind, with the reason. Every ✓ cites a platform run or a test in this actor's suite.

What users wantHow often they said itThis actorVerdictProof
It works, and a failure is never reported as success22 complaints ("It is not working"; "0 rows, still SUCCEEDED")A pool Indeed refuses moves to residential IPs by itself; a refused company is error, never "no reviews"✓run 0PIcSeTQwsaMTDgfl (every datacenter IP refused: 20 reviews in 3.5 s), test_refused_calls_are_errors_not_no_reviews
Every review the page shows, provably17 complaints ("the page shows 2,900 reviews, I get 40")Walks to Indeed's own count; complete and stop_reason on every company✓runs W86lrTVRzc12h9w0J (719 of 719), Q04InzAIQ5t6Heyzu (59,236 of 59,237), UScsZyIouoKHVmKWN
Never pay twice for the same reviews6 complaints ("every retry re-buys the whole company")A resurrected, migrated or resumed run bills nothing twice✓runs u4ZUzPwd5Yy7IMaMK, JtMesmY9uv4fWaqRW, 7DvrvzwxI9e08Qlai
Limits that hold: maxItems, maximum cost, Abort10 complaints ("maxItems 20, it returned 3,000 and cost $70")Every cap is exact, and an abort stops the bill at once✓runs H9yZVlxL8P8DvEmGI, emWx0BafTvP4tBFGH, NN68FmbCebu52Rcfh
Filters that work: date, location, a pasted URL's own filters12 complaintssinceDate/untilDate, Indeed's own URL filters kept✓runs FTRWfqHvPrIATQ4dk, UULDcA9JolXCFk9dh, gSMgvBZdpCn1e2uHN
Simple input: names, websites, long lists8 complaints, praise for "simple setup"A name is enough; a company's website works; one bad entry never voids a list✓runs 10uif2asbwFQwsnBh, CowdxWdRHIhVePMQc, WZ8GVuqulFHoNAUxz
No hidden caps: free plan, a single country8 complaintsNo plan check; the run says how many reviews the chosen countries leave out✓run 9S9BHMgTdN2kbg9eP, test_a_run_says_how_many_reviews_the_chosen_countries_leave_out
Output that drops into a spreadsheet or an automation9 complaints ("1,100 columns"), praise for "easy to transform"19 named review columns, ISO dates, one flat line per company✓runs MpaBxwVc1kha0upsL, 8ox5bEXjZVhWjni1V
Speedpraise ("impressively fast"), 4 complaints ("40 minutes to 2 hours")100 reviews: the first after 2.6 s, all in 7.2 s; memo23 5.3 s and 20.7 s on the same 100✓runs As5OCqbxtvCGrK4Ud, oxpSPmgusgbioDPvP
Big employers without timeouts4 complaintsStops before its time limit with what it read and continues with resumeFromRun✓runs emWx0BafTvP4tBFGH, JwVdbYJsPVNrHgz8D
Limit how far back it goespraise (memo23, twice)sinceDate, untilDate, onlyNewReviews✓runs FTRWfqHvPrIATQ4dk, u2Z3hSDtz7hQHTDcI
Help when something goes wrongpraise (12 of 28 five-star reviews), 4 complaintsNo issue has been opened on this actor yet, so there is nothing to compare; every run card names what happened and what to do=test_the_run_card_names_what_was_not_found_with_indeeds_closest_name, test_the_run_card_names_a_failed_company_and_where_to_report_a_failed_run
Interview questions2 complaints, praise ("a working interviews scraper")The Interviews topic of the company's Q&A, with top answers, from 2026-12-02=test_interview_questions_are_the_interviews_topic_of_the_q_and_a
Interview difficulty and experienceasked for in the same threads✗ sign-in wall: Indeed sends its interview pages to the company overview for signed-out visitors✗run zghI36Zj7vvdyCZJ9
Reliability as Apify's run statistics show itpraise ("3,500 runs and no problem")96.6% of 1,753 runs in the 30 days to 2026-11-19, against memo23's 99.8%: all 54 failures come from Indeed's block of 2026-11-06/07, none since the fix of 2026-11-18✗ recovers with time: those runs leave the 30-day window by 2026-12-07Apify's public run statistics
Price4 complaints, praise ("pricing is good")$3 per 1,000 reviews, no start fee; memo23 $1.50 plus $0.005 a run✗ price: the owner's decisionthe Pricing tab

Remaining ✗: price (the owner's decision), interview difficulty (sign-in wall), the 30-day success rate (recovers by 2026-12-07); in the features table also WhatsApp and MCP-connector delivery (not built: WhatsApp needs your own Meta Business account and phone number, which cannot be tested here, and Apify's connector delivery has no public API) and job rows (built and measured, waiting for their price).

How it compares with the other Indeed company scrapers

The five Indeed company-review scrapers with the most users in the 30 days to 2026-11-19, from their own input forms and READMEs, against this actor. Every ✓ of this actor cites a test in its suite or a platform run; the rivals' cells are what their forms and READMEs offer. The last column is the audit's verdict: ✓ better than every rival, = as good as the best of them, ✗ behind, and why. Tally: 15 ✓, 9 =, 4 ✗.

This actormemo23 (100 users, two listings)kaix company profiles (9)reviewly (5)silentflow (5)blackfalcondata (3)Verdict
Company names, not only URLs✓ test_a_plain_company_name_is_read_under_indeeds_own_slug✗ URLs✓✗ URLs✓✓=
A company's own website as input✓ run CowdxWdRHIhVePMQc✗✗✗✗✗✓
Find the employers by keyword and place✓ run XB1qXnZeOfwcukt6i (trucking in Texas: 50 listed), test_a_search_skips_the_namesake_indeed_lists_first✗✓ profiles only✗✗✗✓
Every country in one walk, or any one country✓ run W86lrTVRzc12h9w0J (719 of 719 worldwide)one site per URLone site✗one site✗✓
A cap per company and one for the run✓ runs H9yZVlxL8P8DvEmGI, JGO9eKW66CcHdT9LQper URL–per companyper company✓=
Reviews since a date✓ run NrMUdgRqRaNTguj1c✓–✓✗✗=
Reviews up to a date (a date window)✓ run u2Z3hSDtz7hQHTDcI (December 2025 only: 167 newer reviews dropped, free)✗–✗✗✗✓
Lowest or highest rated first, a rating range✓ test_rating_filters_run_before_billing✗–✗✗order only✓
Indeed's topic filter (management, pay …)✓ run RPaY1jAHPbMN10MaX (Google, US, management: 994 reviews)✗–✗✗✗✓
One language✓ run IAJUHwgdW5cCXfrXL (Spanish, every country)✗–✗✗✗✓
Job title, location, keyword, current or former✓ test_text_job_title_location_and_since_filters, test_keyword_and_employment_status_filters_run_before_billing✗–✗✗✗✓
Only new reviews on a schedule✓ test_only_new_reviews_delivers_and_bills_a_review_once_across_runs✓✗✗✗✓=
Separate memories for two monitors✓ test_state_key_keeps_two_monitors_of_one_company_apart, run w0c2I28PtgrbRIWn5✗✗✗✗✓=
Alerts to Slack, Discord, Telegram or a webhook, with headers✓ test_an_alert_carries_what_the_run_delivered_in_the_shape_each_service_wants, run UFSLlXzUCdTLpShr6 (a header-carrying POST answered 200)✗✗✗✗✓=
WhatsApp alerts, delivery through Apify MCP connectors✗✗✗✗✗✓✗ not built: WhatsApp needs your own Meta Business account and phone number, which cannot be tested here; connector delivery has no public API
Compact rows and empty fields left out, for AI agents✓ run JAwoSe9a5szUHqD6i, test_compact_and_exclude_empty_fields_shape_rows_never_the_bill✗✗✗✗✓=
Company profile: ratings, size, CEO, happiness, open jobs✓ free, test_company_and_review_rows_carry_the_free_extras✓✓ its product✗✓✓=
How many employees rated the CEO, Indeed's grade for each happiness score, the pay poll✓ free (1.18), run nH9sfT7V1SuuurSkE (Google: 66% of 2,104; Starbucks' grades from POOR to GOOD), test_the_company_row_carries_the_ceo_rating_count_the_happiness_grades_and_the_pros_cons_survey_count✗✗✗✗✗✓
Salaries, Q&A, interview questionsfrom 2026-12-02: test_salary_and_question_rows_are_charged_exactly_as_delivered✓✗✗✓interviews= from 2026-12-02
Interview difficulty and process figures✗ not shown to signed-out visitors (run zghI36Zj7vvdyCZJ9)claimed✗✗✗claimed✗ sign-in wall
Open job listings, with each job's description, benefits and shiftsbuilt and measured (runs PEZdmUBMBlFwCnMCk, 8Ljh1bawIz6X4i9MR), waiting for its price (gated)✓ with descriptions✗✗✓✗✗ price: the owner's decision
Business contacts from the company's websitefrom 2026-12-02, emails, phones and social profiles: test_contacts_found_are_billed_once_per_company_on_the_company_row✓ emails✗✗✗✗✓ from 2026-12-02
A cut-short run continues without paying twice✓ runs u4ZUzPwd5Yy7IMaMK, 7DvrvzwxI9e08Qlai✗✗✗✗✗✓
complete and stop_reason for every company✓ run UScsZyIouoKHVmKWN✗✗✗✗✗✓
100 reviews of one company: first review / all of them✓ 2.6 s / 7.2 s, run As5OCqbxtvCGrK4Ud5.3 s / 20.7 s, run oxpSPmgusgbioDPvP––––✓
A CSV of reviews only, with named columns and ISO dates✓ run MpaBxwVc1kha0upsL (17 lines, 17 reviews)one review a line, raw columns (title/text, "November 16, 2026"), run oxpSPmgusgbioDPvP––––✓
A company table that opens in a spreadsheet✓ one flat line per company, run 8ox5bEXjZVhWjni1V–✗ about 1,100 columns ("Nobody needs or can work with 1100 columns", its issue of 2026-05-20)–––✓
Price per 1,000 reviews$3, no start or run fee$1.50 + $0.005 a run(company profiles only)$8 + $0.50 a run$2.50$1.50 + $0.001 a run✗ price: the owner's decision

What users praise in other company-review scrapers, measured on this one

The 29 four- and five-star text reviews of the Indeed and Glassdoor review scrapers on Apify Store (read 2026-11-19), grouped by what they praise, with the same job measured on this actor. Runs are on its live build of 2026-11-19 unless noted.

Users praiseInThis actor, measuredProof
Speed ("impressively fast", "results within 30 seconds each run")memo23's Glassdoor and Indeed scrapers100 Google reviews: the first after 2.6 s, all 100 in 7.2 s, where memo23's Indeed reviews scraper took 5.3 s and 20.7 s for the same 100 (2026-11-17). 200 Home Depot reviews from every country in 11.7 sruns As5OCqbxtvCGrK4Ud, oxpSPmgusgbioDPvP (memo23), 8GHz0SN8OBSUAUutp
Simple setup ("great for beginners", "the setup is simple")getdataforme, memo23a company name is enough, where memo23 asks for Indeed URLsrun 10uif2asbwFQwsnBh ({"companies": ["Home Depot"]}: the 100 newest reviews of The-Home-Depot)
Accurate, complete datamemo23every review Indeed declares, each once: Nucor 719 of 719 worldwide, Starbucks 59,236 of its 59,237 US reviewsruns W86lrTVRzc12h9w0J, Q04InzAIQ5t6Heyzu
Output that drops into dashboards and automations ("clear output that's been easy to transform", "easy to integrate into my automations")memo23, getdataformeflat rows, named fields, YYYY-MM-DD dates; a CSV of reviews only with 19 named columns (statusRows: "none", includeCompanyRow: false) and a Companies view with one flat line per company. memo23's CSV of 100 reviews has one review per line too, but 42 raw columns such as title/text and compensationAndBenefitsRating/type, and dates like "November 16, 2026"runs MpaBxwVc1kha0upsL (17 reviews, 17 lines), 8ox5bEXjZVhWjni1V (Companies view: 59 columns, 649 on 1.16 in run vnQzMEmDwygv8Edmu), oxpSPmgusgbioDPvP (memo23); tests test_status_rows_none_keeps_the_dataset_to_reviews_and_the_status_record_lists_every_company, test_every_view_has_readable_labels_and_the_companies_view_exports_one_flat_line_per_company
Limiting how far back it goes ("to minimise the budget used")memo23 (two reviews)sinceDate ends the walk at the date and drops the older reviews on its last page free: Google's last 14 days, 5 reviews delivered, the 35 older ones read on the way not charged, complete: true. untilDate closes the window; onlyNewReviews takes what is new since the last runrun FTRWfqHvPrIATQ4dk
Reliability ("over 3,500 runs so far and no problem")memo231,753 runs in the 30 days to 2026-11-19, 96.6% succeeded: all 54 failures came from Indeed's block of 2026-11-06/07, none since the fix of 2026-11-18. A proxy pool that Indeed refuses moves to residential IPs by itself: 20 reviews in 3.5 sApify's public run statistics; run 0PIcSeTQwsaMTDgfl
A developer who answers fast (12 of the 28 five-star reviews)memo23no issue has been opened on this actor yet. Every run says, per company, what happened and what to do: a name not found comes with Indeed's closest name, a failure with what was charged (nothing) and where to report ittests test_the_run_card_names_what_was_not_found_with_indeeds_closest_name, test_the_run_card_names_a_failed_company_and_where_to_report_a_failed_run
Price ("pricing is good compared to other Glassdoor actors")kaix, scrapifier$3 per 1,000 reviews, no start fee; filtered, empty and not-found results are free. memo23's Indeed reviews scraper: $1.50 per 1,000 plus $0.005 a runthe Pricing tab
Interview data ("been looking for a working interviews scraper for so long")scrapifier (Glassdoor)interview questions with their top answers, from 2026-12-02; Indeed shows signed-out visitors no interview difficulty or experiencetest test_interview_questions_are_the_interviews_topic_of_the_q_and_a

Where other scrapers fail, and what this actor does instead

What users of other Indeed and Glassdoor review scrapers report in public issues and reviews, and what this actor does about each. Every line is backed by a test in this actor's suite (tests/test_actor_indeed_reviews.py, tests/test_indeed.py) or a platform run.

Where others failWhat this actor doesProof
Fewer reviews than the page shows ("the page shows almost 2,900 reviews but I get 40"; a run that "succeeded" after stopping mid-company)Walks to the count Indeed declares and never bills Indeed's silent repeat of page one past the end; a page refused half-way is reported as partial, never as the end. When the countries you chose hold at most half of a company's reviews, the run card says how many it left out and how to get them.test_the_wrap_around_never_bills_page_one_twice, test_a_refusal_half_way_keeps_and_bills_what_was_delivered, test_a_run_says_how_many_reviews_the_chosen_countries_leave_out; run 9S9BHMgTdN2kbg9eP (Tesco under us: "1,485 of 1,619 reviews of Tesco are outside the countries you chose"), run Q04InzAIQ5t6Heyzu (Starbucks' 59,237 US reviews: 59,236 delivered, each once, complete: true)
Filters in a pasted URL ignored (a location or country filter dropped; "909 duplicates instead of around 200 relevant results found via link")A pasted review URL is read as Indeed serves it: its country (fcountry or the host), city (floc), job title (fjobtitle), text search (ftext) and order (sort)test_a_pasted_review_url_keeps_its_own_filters_and_order, test_a_pasted_url_with_indeeds_country_filter_is_read_for_that_country; runs gSMgvBZdpCn1e2uHN (2026-11-19: Google's Software Engineers, 20 of 20 with Indeed's own job title), a7xyJgqRHc9zxkjvp (Software Engineers, lowest rated first), UULDcA9JolXCFk9dh (a uk.indeed.com URL with floc=London: London reviews only), cQgbvSm9xKodk8WV4 (?fcountry=GB: UK reviews)
maxItems ignored, an uncapped billmaxItems is one exact total for the run, and the run's maximum cost is never passed, over every paid row typetest_max_items_caps_the_total_and_never_reads_the_next_company, test_one_spend_cap_over_four_events_is_never_exceeded_and_charged_equals_delivered; runs 5VFhDzFxFWIl6PnvI (maxItems 30: 30), emWx0BafTvP4tBFGH (a $20.2516 cap met with 6,750 reviews, the last page charged 10 of 20)
Date filter broken ("I'm getting all results from the beginning of time")sinceDate takes a date or a window like 30 days and ends the walk there; other scrapers' targetDate and since are read as ittest_text_job_title_location_and_since_filters, test_rival_settings_keep_their_meaning; run FTRWfqHvPrIATQ4dk (2026-11-19, Google's last 14 days: 5 reviews, the 35 older ones on the way dropped free, complete: true)
The same review billed twice (two links to one company)A company given twice is read once, and a review two targets share (a name and its worldwide URL) is billed oncetest_a_name_and_its_slug_are_one_company_billed_once, test_a_name_and_its_worldwide_url_are_two_targets_that_bill_a_review_once; run RePkM7r7cQtFI0ehd (56 reviews from a name and its worldwide URL, none twice)
Blocked, "not working", a 403Keeps the proxy IPs Indeed serves and starts the next run on them, spreads each sweep over three browser fingerprints, falls back to residential IPs only when the whole pool is refused and moves back when an IP answerstest_the_next_run_starts_on_the_datacenter_session_indeed_served_before, test_a_wave_spreads_over_three_profiles_and_the_one_served_becomes_the_runs, test_a_walk_on_residential_moves_back_to_the_datacenter_when_an_ip_answers; probe runs hFSTJ5tIenwBz3uy6, 88X8UgTfrgEOR4tmr (2026-11-18: 1 of 27 datacenter IPs served); run 0PIcSeTQwsaMTDgfl (2026-11-19, a proxy pool that refuses everything: 20 reviews in 3.5 s over residential IPs)
A failure reported as successA company Indeed refused is error, never no_reviews; one refused half-way is partial with what was read delivered and billedtest_refused_calls_are_errors_not_no_reviews
Timeouts on big employersStops before its own time limit with what it read, names the run to continue from, and continues with resumeFromRun; a walk over 40,000 reviews logs its pace and the time it needstest_a_run_cut_by_its_time_limit_can_be_resumed_where_it_stopped, test_a_whale_walk_logs_its_pace_and_when_the_time_limit_comes_first; runs emWx0BafTvP4tBFGH (6,750 Walmart reviews over 338 pages, no page lost) and JwVdbYJsPVNrHgz8D (a 2-minute run: the pace line, then green at its time limit with a resume point)
No interview dataInterview questions with their top answers, from the Interviews topic of the company's Q&A (from 2026-12-02)test_interview_questions_are_q_and_a_rows_switched_on_like_the_q_and_a_add_on, test_interview_questions_are_the_interviews_topic_of_the_q_and_a
Billed twice after a retry ("there is no resume, so every retry re-buys the whole company"; "when it resurrects, it keeps doing the same data over and over again")A resurrected or migrated run reads back what it delivered and bills none of it again; a run stopped by its time limit, its maximum cost, an abort or a refusal names the run to continue from with resumeFromRuntest_a_resurrected_run_bills_no_review_twice_and_reads_no_finished_company_again, test_a_run_killed_between_two_pages_carries_on_after_a_restart_and_bills_each_review_once, test_a_spending_limit_stop_says_how_far_back_it_got_and_resume_from_run_carries_on; runs u4ZUzPwd5Yy7IMaMK (a finished run resurrected: 4 unique, 4 charged, no row twice), JtMesmY9uv4fWaqRW (aborted at once at 180 reviews, resurrected to 300 unique = 300 charged), 7DvrvzwxI9e08Qlai (continued eiT38iemZuclZPTIS after its spending limit: 0 overlap)
A truncated dataset that looks complete ("a truncated dataset is indistinguishable from a complete one")Every status row says complete, stop_reason and oldest_review_date; the run card counts the complete companies; failOnIncomplete ends a short run FAILEDtest_every_status_row_says_whether_the_company_is_complete_why_it_stopped_and_how_far_back_it_got, test_fail_on_incomplete_fails_only_for_stops_the_input_did_not_ask_for; runs UScsZyIouoKHVmKWN (Nucor worldwide: 719 reviews, complete: true, end), BaINNHoh9tO0gvUmQ (failOnIncomplete with a name Indeed does not know: FAILED, the 3 reviews kept)
One bad link fails the whole listThe bad entry gets a free invalid_input status row and the rest of the list runstest_one_bad_entry_gets_a_free_status_row_and_the_rest_of_the_list_runs; run WZ8GVuqulFHoNAUxz (the input that ended "Input problem" in run BO4wktzjdIVt6av1Q: now 4 reviews and one free invalid_input row)
Only Indeed links accepted ("I have a list of company names, I don't have the URLs"; lead lists of company websites)Names, slugs and URLs as before, and now a company's own website: read under the Indeed company whose own page lists that site, never a near misstest_a_company_website_is_read_under_the_indeed_company_that_lists_it, test_a_website_no_indeed_company_lists_is_not_found_and_free_never_a_near_miss; run CowdxWdRHIhVePMQc (Lacoste, HubSpot and Zendesk by their home pages; starbucks.com, which Starbucks does not list on Indeed, came back with suggestions, free)
A CSV with a thousand columns ("Nobody needs or can work with 1100 columns")The Reviews view exports the 19 review columns, and with statusRows: "none" and includeCompanyRow: false every line of it is a review. The Companies view exports one flat line per company: rating, counts, histogram, category averages, happiness, profile; 1.16 flattened its job, competitor and Q&A lists into 649 to 723 columnsruns e2lhXxxI8PmxFLkh5 (19 columns, against 987 without the view in run X6oDH8IMJdfAilihc), MpaBxwVc1kha0upsL (reviews only: 17 lines for 17 reviews), 8ox5bEXjZVhWjni1V (Companies view: 59 columns); tests test_status_rows_none_keeps_the_dataset_to_reviews_and_the_status_record_lists_every_company, test_every_view_has_readable_labels_and_the_companies_view_exports_one_flat_line_per_company
No answer, no way to get help ("Could not find a contact email to address the problems", "Does not respond to support requests")Every run says what happened to each company and what to do: a name Indeed does not know comes back with Indeed's closest name, a refused company with what was charged for it (nothing) and how to retry, a run where every company failed with where to report it: the Issues tab, with the run idtest_the_run_card_names_what_was_not_found_with_indeeds_closest_name, test_the_run_card_names_a_failed_company_and_where_to_report_a_failed_run, test_the_run_card_names_a_company_without_reviews

Questions users ask about other scrapers

The questions below come from the public reviews and issues of other Indeed and Glassdoor review scrapers (2026-11-18). Each answer cites a test in this actor's suite or a platform run.

Is the free plan limited?

No. The actor has no plan check: a free-plan account gets the same walk, the same caps and the same speed, and a run ends only where your caps or your maximum cost say. A free account's proxy pool is small (5 datacenter IPs); when Indeed refuses all of them the run moves to residential IPs by itself — with the pool forced to refuse everything, 20 reviews in 4 seconds (run 5I7oEKcwCNhDNtBzc).

Will I get every review the page shows?

Set countries: ["all"] and maxReviewsPerCompany: 0. Nucor gave 719 of the 719 reviews Indeed declares, 719 unique and 719 charged (run W86lrTVRzc12h9w0J), and on 1.13 its status row says complete: true, stop_reason: "end" (run UScsZyIouoKHVmKWN, 719 again). Past 40,000: up to 1.13 every walk stopped at 2,000 pages, so Starbucks' 59,237 US reviews gave 40,000 with stop_reason: "max_pages" (run Q04InzAIQ5t6Heyzu); the same run resurrected on 1.14 read back its 40,000 and finished: 59,236 reviews, each once, complete: true (a private copy of this build's source, 2026-11-18). Under the default us the run card says how many reviews other countries hold: "1,485 of 1,619 reviews of Tesco are outside the countries you chose" (run DOh2jZ5CDbhpWkJ1R).

How do I know a dataset is complete and not cut short?

Every status row carries complete, stop_reason and oldest_review_date; SUMMARY counts them, and the run card ends complete: N of M companies with why the others stopped (stopped at maxReviewsPerCompany, cut by the spending limit, …). With failOnIncomplete: true a run cut by something you did not ask for ends FAILED, with what it delivered kept (run BaINNHoh9tO0gvUmQ). Tests: test_every_status_row_says_whether_the_company_is_complete_why_it_stopped_and_how_far_back_it_got, test_fail_on_incomplete_fails_only_for_stops_the_input_did_not_ask_for.

Do maxItems, the maximum cost and Abort really stop the bill?

Yes. Another scraper's maxReviews: 10 gave 10 (run H9yZVlxL8P8DvEmGI); memo23's bulk shape, 3 URLs with maxItems: 5, gave 5 per URL (run DOh2jZ5CDbhpWkJ1R); maxItems: 30 gave 30 (run JGO9eKW66CcHdT9LQ); the maximum cost is met exactly and the run card says how far back it got (run bWTdc4rKQFVhEw853: "spending limit reached (Google read back to 2026-11-10)"); an abort ended a run 2.2 seconds later with 260 reviews delivered and 260 charged (run NN68FmbCebu52Rcfh). Rows a cap or the maximum cost refused are neither delivered nor charged.

If I resurrect or resume a run, or Apify migrates it, am I billed twice?

No. Version 1.12 re-billed a resurrected run (8 reviews charged for 5 unique, run 0rQQ87K7yNJIbiLiW). From 1.13 a run the platform starts again reads back what it delivered: a finished run resurrected bills nothing more and writes no row twice (run u4ZUzPwd5Yy7IMaMK: 4 unique, 4 charged), and a Walmart walk aborted at once at 180 reviews, then resurrected, carried on to 300 unique reviews with 300 charged (run JtMesmY9uv4fWaqRW). A run stopped at its maximum cost names the run to continue from, and the continued run overlapped it by 0 reviews (run 7DvrvzwxI9e08Qlai after eiT38iemZuclZPTIS); so did a run continued after a graceful abort (run dvoabVYDeR9GPRYHc after 6oFkN4AdiVDmcgHtz). A run aborted at once saves no resume point, and a run may not read another run's storages, so resumeFromRun of it says to resurrect it instead and charges nothing (run fP5bk6VbkhLht8xdG). With statusRows "none" or "problems" the status rows are not all in the dataset, so a restarted run reads them from its STATUS record: a finished run of that kind, resurrected, read no page and delivered and charged nothing more (run Xb1zis03DEiO8wgZz, test test_a_restarted_run_with_status_rows_none_reads_its_status_record_and_no_finished_company_again). Tests: test_a_resurrected_run_bills_no_review_twice_and_reads_no_finished_company_again, test_a_run_killed_between_two_pages_carries_on_after_a_restart_and_bills_each_review_once, test_a_graceful_abort_stops_at_the_next_page_and_saves_where_each_company_stopped.

No. It gets a free invalid_input status row and the rest of the list runs: Google, a Glassdoor link and Starbucks gave the Google and Starbucks reviews plus one free row for the link (run WZ8GVuqulFHoNAUxz); version 1.12 ended that input "Input problem" and read nothing (run BO4wktzjdIVt6av1Q).

Do I get interview difficulty and experience?

No scraper can signed out: Indeed removed its interviews tab (/interviews redirects to the company overview) and shows no difficulty or experience to visitors. What it does publish are the interview questions in the Interviews topic of the company's Q&A: from 2026-12-02 includeInterviews delivers them with their top answers; until then an /interviews URL gives the company's reviews with a note (run zghI36Zj7vvdyCZJ9).

Can I get reviews in one language?

Paste the review URL with Indeed's own language filter: https://www.indeed.com/cmp/Google/reviews?fcountry=ALL&lang=es gave Spanish reviews filed under Spain, Mexico and Chile (run yceQeZ3a90Km0sCp9). Indeed does not count reviews per language — it lists the whole company's 6,297 for that view — so from 1.14 its status row leaves found_review_count empty and says why in note (run 46zrtmYc4sZEHSS3z).

How long does a big company take?

About 1,000 to 1,300 reviews a minute on the IPs Indeed serves: 40,000 Starbucks reviews in 30.5 minutes (1,311 a minute, run Q04InzAIQ5t6Heyzu), 6,750 Walmart reviews in 6.9 minutes (run emWx0BafTvP4tBFGH): 10,000 reviews in about 8 to 10 minutes, where users of one rival report 40 minutes to 2 hours for as many. A run stops before its own time limit with what it read and continues with resumeFromRun.

I have a list of company websites, not names. Does that work?

Yes. Give the websites as they are (https://www.lacoste.com, www.apple.com). Each is read under the Indeed company whose own Indeed page lists that site, so a site never buys another employer's reviews; when none of the companies Indeed suggests lists it, the entry is not_found and free, with those companies and the sites they list in suggestions. Tried on 15 home pages on 2026-11-18, 10 resolved exactly (Lacoste, Apple, HubSpot, Patagonia, Zendesk, Nucor, Etsy, Dropbox, Gusto, Chipotle); Basecamp and The Home Depot were not among the five companies Indeed suggested for their names, Tesco and Starbucks list no website on Indeed, and Spotify lists its jobs site, so those five came back with suggestions instead: give their names or Indeed URLs. Run CowdxWdRHIhVePMQc (Lacoste, HubSpot and Zendesk by their home pages; starbucks.com, which Starbucks does not list on Indeed, came back not found with suggestions, free) (tests test_a_company_website_is_read_under_the_indeed_company_that_lists_it, test_a_website_no_indeed_company_lists_is_not_found_and_free_never_a_near_miss).

Where do I report a problem?

In this actor's Issues tab, with the run id. Every status row and SUMMARY say what happened to each company and why a walk stopped (test test_every_status_row_says_whether_the_company_is_complete_why_it_stopped_and_how_far_back_it_got), so an answer usually starts there.

Review rows: every field this Indeed scraper returns

Three row types share one dataset and are told apart by type.

type: "review" — charged, $0.003 each

FieldWhat it is
review_idIndeed's stable encrypted review id, also the id in the review URL
companyThe slug Indeed files the company under, the part after /cmp/ — also when you gave a plain name
company_nameCompany name as Indeed displays it
countryWhich domain the row was read on: us, all, ca, de, …
ratingOverall rating, 1-5 stars
work_life_balanceThe reviewer's 1-5 rating for work/life balance, or null if they skipped it
compensation_benefitsTheir 1-5 rating for pay and benefits, or null
job_security_advancementTheir 1-5 rating for job security and advancement, or null
managementTheir 1-5 rating for management, or null
culture_valuesTheir 1-5 rating for culture, or null
titleThe headline the employee wrote
textThe full review body, untruncated
prosThe Pros field, when the review was submitted with the pros/cons form
consThe Cons field, likewise
job_titleThe reviewer's job title
locationWhere they worked, e.g. Mountain View, CA
review_countryISO country code Indeed files the review under
employment_statuscurrent or former
review_datePublication date as YYYY-MM-DD
review_date_textThe date exactly as Indeed printed it, in that domain's language
helpful_countHow many readers voted it helpful
unhelpful_countHow many voted it not helpful
employer_replyThe employer's official public response, if any
employer_reply_dateDate of that response
job_title_normalizedIndeed's normalised form of the job title, for grouping reviews by role
urlDirect link to the review

type: "company" — free, one per company

review_count (all reviews Indeed holds worldwide), found_review_count (what the domain you read actually lists), rating, histogram (how many gave 1, 2, 3, 4 and 5 stars), ratings_breakdown (the company averages for the same five categories), happiness_score and happiness_grade (Indeed's Work Happiness measure), top_pros and top_cons (what Indeed extracts from the corpus), topics (name, rating and count per theme), industry, company_size, headquarters, website, founded, revenue, ceo, reviews_fetched and company_url. From 1.7 also (the last four rows from 1.18, read off the same pages, free):

FieldWhat it is
ceo_approvalShare of employees who approve of the CEO, in percent
rating_by_yearThe average rating of each of the last five years' reviews, e.g. {"2025": 4.31, "2026": 4.21}
happiness_scores, happiness_responsesIndeed's fifteen work-happiness scores (0-100: Happiness, Appreciation, Learning, Flexibility, Compensation, Stress-free, …) and how many answers they rest on
reviews_by_job_titleThe job titles with the most reviews, with their count and average rating
reviews_by_locationThe locations with the most reviews, with their count and average rating
reviews_by_countryReviews per country they are filed under, e.g. {"US": 4021, "IN": 643, "GB": 212} — any of these codes works in countries
open_jobs, open_jobs_by_location, open_jobs_by_titleThe company's open jobs on Indeed, and where and what they are
salary_count, salary_previewHow many salaries are reported, and the profile's salary preview (job title, typical salary, period, reports)
question_countQuestions asked about the company in its Q&A
similar_companies, related_companiesCompetitors Indeed lists next to the company (name, slug, URL, rating, review and salary counts) and its brands and subsidiaries
description, logo_url, employer_id, profile_claimedThe company's own description, its logo, Indeed's stable employer key, and whether the employer manages its profile
similar_toOn a company read through maxSimilarCompanies: the company it is similar to
recent_jobsThe company's most recent open jobs as its profile shows them (title, location, salary as shown and as a range, how long ago it was posted, job types, urgent hire, Indeed Apply, job page)
popular_questionsThe Q&A preview of its profile: question, answer count, top answer, whether the employer wrote it
sectorsThe sectors Indeed files it under
ceo_approval_count, ceo_photo_urlHow many employees rated the CEO, where Indeed's own page states it ("Ratings from 2,104 employees" for Google on 2026-11-19; Starbucks' page states none), and the CEO's photo
happiness_gradesIndeed's own grade for each of the fifteen happiness scores, as its page colours them: EXCELLENT, GREAT, GOOD, FAIR or POOR (Google's Stress-free: FAIR; Starbucks' Achievement: POOR)
pros_cons_responsesHow many survey answers Indeed's top_pros and top_cons stand on (Google: 625, Starbucks: 20,409)
salary_satisfaction, salary_satisfaction_responsesWith the salary add-on, from the salaries page it reads anyway: the share of employees who say their pay covers the cost of living where they live, and how many answered (Google: 66% of 485)

type: "status" — free, one per company and country

target (what you typed), company (the slug it resolved to), company_name, country, status (ok, no_reviews, not_found, duplicate, partial, limit, done, invalid_input or error), reviews delivered and charged, complete (true when nothing your filters and date limit let through was left unread), stop_reason (end, since_date, rating_bound or only_new when complete; cap, max_items, spending_limit, time_limit, aborted, refused, not_found or invalid_input when not), oldest_review_date (how far back the walk got), filtered dropped before billing, already_seen skipped by onlyNewReviews, pages read, total reviews Indeed holds worldwide, found_review_count, reviews_outside_countries (when the countries you chose hold at most half of the company's reviews: how many they left out), duplicates, error, suggestions (on not_found: the closest companies Indeed knows, each with name, company and url) and note (what was done with a sub-page URL). This row is the contract: if a company could not be read, you get status: "error" with the reason, never a silent empty result that looks like a company with no reviews; if Indeed refused a page half-way, you get partial and keep every review delivered before it. statusRows decides which of these rows go into the dataset; with problems or none all of them are in the run's STATUS record. The run also writes a SUMMARY record to the key-value store with the same counts for the whole run, plus complete (every company complete), complete_companies, incomplete (how many stopped for each reason) and oldest_review_date, and the run card ends with complete: N of M companies and why the others stopped.

The dataset ships with eight views in the Output tab, each with readable column names: Overview, Reviews, Complaints (rating, cons, management and work-life-balance scores, and the employer reply side by side), Companies (one flat line per company: rating, review counts, the star histogram, category averages, happiness, industry, size, headquarters, website, CEO; the long lists of jobs, competitors, Q&A and reviews per country stay in the JSON), Salaries, Q&A, Jobs and Status. A view picks columns, not rows: exported, the Reviews view also has a line for each company row and status row. statusRows: "none" keeps the status rows out of the dataset (they go to the run's STATUS record) and includeCompanyRow: false the company rows, so the export is reviews only; statusRows: "problems" keeps just the status rows with something to say.

Company reviews across 44 Indeed country domains

Indeed does not serve one global review list. www.indeed.com shows the reviews filed under the United States — 4,050 of Google's 6,270 on the day this was verified, but only 130 of Tesco's 1,596 — while uk.indeed.com shows the ones filed under the United Kingdom (218 for Google, 1,365 for Tesco) and de.indeed.com the German ones, each in that country's language. Every row says which in review_country. The subsets barely overlap.

So countries is a multiplier, and it has one special value:

  • "us" — www.indeed.com, the reviews filed under the United States. The default.
  • "all" — www.indeed.com with Indeed's own worldwide filter. For Google that is 6,270 of 6,270 reviews in a single walk. If you want everything for a company, or the company is not American, this is the value to pass.
  • Any of ca, uk, ie, au, nz, in, sg, ph, za, ae, hk, id, th, vn, lu, ch, be, de, at, fr, nl, es, it, pt, br, mx, ar, cl, co, pe, pl, se, dk, no, fi, tr, gr, hu, cz, ro, ua, jp, tw — that country's own subset, in that country's language.
  • Any other two-letter country code (my, kr, eg, ng, pk, il, sa, …) — read on www.indeed.com through Indeed's own country filter: the reviews filed under that country (Google, Malaysia: 27 on 2026-11-17). The free company row's reviews_by_country says which countries a company has reviews in.

The run tells you when the default misses most of a company. Under the default us, Tesco gives 134 of its 1,619 reviews (2026-11-18). Such a run now says so: the company's status row carries reviews_outside_countries: 1485 and a note, SUMMARY lists it under outside_countries, and the run card ends 1,485 of 1,619 reviews of Tesco are outside the countries you chose (us); set countries: ["all"] to read them. It speaks only when the countries chosen hold at most half of a company's reviews (Google's US share, 65%, gets no note) and reads Indeed's own per-country counts off the first page, so it costs no extra request. The default stays us so saved inputs and monitors keep working.

A pasted URL names its own country: uk.indeed.com/cmp/…, or Indeed's filter ?fcountry=GB on www, is read for that country whatever countries says.

List several and they are walked in order until the cap is met. A review that appears on two domains is delivered and charged once: the run keeps one set of review ids per company. Dates are converted to YYYY-MM-DD whatever language they arrive in — 3. September 2025, 2026, augusztus 25. and 2026年2月27日 all resolve — and the original string is kept in review_date_text so nothing is lost.

Employee reviews for employer brand and HR analytics

The five sub-ratings are what make this dataset useful rather than merely large. An overall score of 3.8 says nothing actionable; management: 2 against compensation_benefits: 4 says exactly where the problem is, and job_title plus location says for whom and where.

A few patterns that work well:

  • Category deltas over time. Average each sub-rating by month: management sliding while pay holds is a leadership problem, not a budget one.
  • Role and site segmentation. jobTitleContains and locationContains isolate the staff you actually recruit, at the plant, store or region where feedback becomes actionable.
  • Competitive employer brand. Compare competitors' histograms (maxSimilarCompanies adds them), not headline scores.
  • Response coverage. The share of negative reviews with an employer_reply is a cheap, honest employer-brand metric.
  • Employee sentiment analysis. text, pros and cons are full length, which is what a topic model or an LLM needs.

Why is this a Glassdoor alternative?

Indeed and Glassdoor publish the same kind of workplace review data: an overall star rating, sub-ratings for pay, management, culture and work-life balance, pros and cons, the reviewer's job title and whether they still work there, and the employer's public reply. Indeed's corpus is far larger for hourly and frontline roles — retail, logistics, healthcare, hospitality — and it is published on plain public pages any visitor can read without an account, which is what lets this actor collect it at volume, cheaply and predictably.

This actor reads Indeed only; it returns no Glassdoor data. For most employer-brand, recruitment and market-research questions Indeed company data answers on its own, and the same company is usually reviewed on both. If you need consumer-side reputation as well, pair this with the Trustpilot Reviews Scraper; if your product is an app, the App Store Reviews Scraper covers the other side of the same story.

Add-ons: salaries, Q&A, interview questions and business contacts (from 2026-12-02)

Three paid add-ons, each off by default. Their prices take effect on 2026-12-02 (Apify's notice period for actors with paying users); until then a run that asks for one skips it, charges nothing for it and says so on its run card and in SUMMARY addons_inactive.

type: "salary" — $0.002 per row (includeSalaries, or a /salaries URL). One row per job title with at least one reported salary: job_title, category, salary (a number) and salary_text (as Indeed shows it, e.g. $162,673), period (YEARLY, MONTHLY, HOURLY, …), currency, reported (how many salaries employees reported for the title) and the title's salary url. Indeed's estimates for titles nobody reported are never delivered. A small company lists every title on one page; a big one is read one job category at a time, the categories with the most reported salaries first, until maxSalariesPerCompany (default 100; Google has about 300 titles). Salaries are read on the first country in countries (the US list for all).

type: "question" — $0.002 per row (includeQuestions, or a /faq URL). One row per question asked about working at the company: question, asked_date, answer_count, the top answer with answer_date, answer_job_title, answer_location, answer_by_employer (the employer's official answer) and answerer_current_employee, and the question's url. In Indeed's order, ten a page, until maxQuestionsPerCompany (default 50).

Interview questions — Q&A rows, $0.002 each (includeInterviews, or an /interviews URL). Indeed sends its old interview pages to the company overview (/cmp/Google/interviews did on 2026-11-18) and keeps interview questions in the Interviews topic of the company's Q&A, /cmp/<Company>/faq/interviews (Google: 14 questions): "How should I prepare for an interview at Google?", "What questions did they ask during your interview?", each with its top answer, the answerer's job title and location, and the date. They are Q&A rows with question_topic: "Interviews", under maxQuestionsPerCompany; a page Indeed answers with the whole Q&A instead is never delivered as interview questions.

Job rows — coming (includeJobs, maxJobsPerCompany, or a /jobs URL). One row per open job: title, location, salary_text with salary_min, salary_max and salary_period, indeed_apply, sponsored and the job page url, and from 1.18 what each job's own page says, read with one more request a job: the full description (plain text, and description_html with Indeed's own headings and lists), benefits ("Health insurance", "401(k) matching" …), shifts ("Day shift", "Weekends as needed" …), attributes (Indeed's labels, such as {"Work setting": ["Coffee shop"], "Encouraged to apply": ["Fair chance"]}), posted ("28 days ago"), job_types, urgent_hire, expired, language, and the pay as a range where the jobs list showed none. The job page is read on Indeed's job search with the job opened, which robots.txt leaves open to every visitor (/viewjob is closed to crawlers and answers a login gate); details_status says read, or why not (refused, not_shown for a job Indeed no longer shows), and a job page is never read for a row your maximum cost would not let through. Measured on Apify's datacenter IPs on 2026-11-19: 20 Google jobs, 20 descriptions, in 29 s (run PEZdmUBMBlFwCnMCk); Starbucks, Basecamp and Google's UK site, 17 of 17 (run 8Ljh1bawIz6X4i9MR); on the residential fallback, 5 of 5 (run KWf7HoA6zGxU4u0Ps). memo23's includeMoreJobDetails asks for exactly this. Built and tested, but its price is not set yet: until it is, a run that asks for job rows skips them, charges nothing for them and says so; the company row's free recent_jobs has the latest twenty meanwhile.

Business contacts — $0.003 per company with a find (includeContacts). The actor opens the website the company's Indeed profile links to (following short links), honours its robots.txt, and reads the home page plus up to three contact or about pages for the business contact details the company publishes there: contact_emails (each with type — role for addresses like info@ or careers@, named for an address that names a person — and the page it was found on), contact_phones (from tel: links and the site's schema.org data, never loose digits), contact_page_url and the social profiles it links (contact_linkedin, contact_facebook, contact_instagram, contact_x, contact_youtube). They are added to the company's row, which is written even when includeCompanyRow is off. It is charged once per company, only when at least one email address or phone number was found; contact_status says what happened otherwise (none_found, no_website, site_unreachable, blocked_by_robots, social_only).

How much does it cost?

$0.003 per review row, and nothing else. There is no monthly rental and no per-run fee.

You wantInputReviewsCost
A quick look at one companymaxReviewsPerCompany: 100100$0.30
One company, everything the US domain listscountries: ["us"], maxReviewsPerCompany: 0~4,000~$12.00
One company, everything worldwidecountries: ["all"], maxReviewsPerCompany: 0~6,250~$18.75
Daily monitoring of 20 employersonlyNewReviews: true on a schedule~40 new a day$0.12 a day
A complaints-only feedsort: "rating_asc", maxRating: 2, maxReviewsPerCompany: 200200$0.60
Salaries of one big employer (from 2026-12-02)includeSalaries: true, maxSalariesPerCompany: 0, maxReviewsPerCompany: 1~300 salary rows + 1 review~$0.60
Contacts for 100 employers (from 2026-12-02)includeContacts: true, maxReviewsPerCompany: 1100 reviews + contacts where found$0.30 + $0.003 per company with a find

Free, always: the company row, the status rows, a company that has no reviews, an unknown company, and every review a filter dropped before billing. The run bills for exactly the review rows it delivers — that invariant is asserted in the test suite across every input mode and every spending limit.

Integrations: n8n, Make, MCP and AI agents

Each recipe runs the actor (or a saved task of it) and reads its dataset. Filter on type = review, unless the run used statusRows: "none" and includeCompanyRow: false: then every row is a review.

  • Make — the Apify app's Run an Actor module (kestrel/indeed-company-reviews, your input as JSON), then Get Dataset Items from the run's dataset, a filter on type, and Google Sheets, Airtable, Slack or a database. Map company_name, rating, title, text, pros, cons, job_title, location, review_date and url.
  • n8n — the Apify node, Run actor and get dataset (or an HTTP Request node with the curl call of the quick start), an IF node on type, then your sheet, CRM or chat. With a Schedule Trigger and onlyNewReviews: true, each execution passes on only the reviews written since the last one.
  • Zapier — the Apify app: run this actor or a saved task, then read the run's dataset items; for a daily feed, start from a finished run of a scheduled task.
  • Google Sheets, with no automation tool — =IMPORTDATA("https://api.apify.com/v2/acts/kestrel~indeed-company-reviews/runs/last/dataset/items?token=<YOUR_APIFY_TOKEN>&format=csv&view=reviews&status=SUCCEEDED") fills a sheet with the reviews of your last successful run, 19 columns. Anyone who can open the sheet can read the token in the formula: keep such a sheet private.
  • Slack, Discord, Telegram or any webhook — notifyWebhookUrl posts what each run delivered when it ends: the count per company and the first reviews with their rating, title and link. With onlyNewReviews on a schedule that is a new-review alert, silent on a quiet day.
  • Schedules — save the input as a task and schedule it in the Console (Schedules): with onlyNewReviews: true every run delivers and bills only what was written since the run before.
  • MCP / AI agents — https://mcp.apify.com?tools=kestrel/indeed-company-reviews, as in the quick start. compact: true keeps rows small for a model's context; the SUMMARY record and the status rows tell the agent what it got and what it did not.
  • Warehouses and BI — the dataset exports as JSON, CSV, Excel, XML or RSS from the same API, and every review row carries every field, so the columns stay the same from run to run.

Short answer: the data is public and collecting it is generally lawful in the EU and the US, but it is personal data about identifiable people, and how you use it is where the legal risk lives. Nothing here is legal advice — take your own.

What we can say plainly:

  • This actor reads only what any visitor can read. No login, no account, no paywall, no private endpoint.
  • Employee reviews are personal opinions written by real people. A job title, a location and a date can identify someone at a small employer even though no name is published: treat every row as personal data under the GDPR and equivalent laws.
  • Do not try to re-identify reviewers. Joining these rows against your HR system, rota or LinkedIn to work out who wrote what is the likeliest way to turn a lawful analysis into an unlawful one.
  • Have a lawful basis and minimise. Aggregate employer-brand analysis is an easier legitimate-interest case than storing individual reviews indefinitely; keep only what you need.
  • Publishing is a separate question. Republishing review text raises copyright and database-right questions too; aggregates and short attributed quotes are far safer.
  • Business contacts are still personal data when they name a person. The contacts add-on returns only what the company publishes on its own website to be contacted. A role address (info@, careers@) identifies the company; a named one identifies a person: have a lawful basis, respect opt-outs and your country's marketing rules (in the EU, e-marketing to individuals generally needs consent).
  • Respect Indeed's terms and be polite. The defaults here are deliberately gentle — one request per second per session — and there is no reason to raise them.

Limits and known traps

  • A page past the end silently repeats page one. Indeed answers ?start= beyond the last page with HTTP 200 and the first twenty reviews again, never an error, so a naive scraper bills the same rows forever. This actor bounds the walk by the review count Indeed declares, the page number it reports and the review ids already seen (the Q&A walk likewise), and charges for nothing extra.
  • Indeed refuses most datacenter IPs, and more of them every month. Its Cloudflare check judges the IP first: with a given fingerprint an address is served every time or never (2026-10-05: 93 of 100 fresh sessions gave the same answer three times running, and a served session answered 125 of 125 consecutive pages). In October 55-70% of datacenter IPs were refused; on 2026-11-18 one of this account's 27 served every consistent browser fingerprint, two or three served now and then and the rest refused all 53, while residential IPs served 5 of 5. So the actor sweeps your account's proxy pool, keeps the sessions that are served for the whole run and remembers them for your next run (an Apify proxy session keeps its IP; key proxy-… in your key-value store indeed-reviews-seen). It falls back to residential IPs only when the whole pool is refused, and while there asks the datacenter pool again in the background, moving back as soon as an IP answers. SUMMARY http counts the pages each served (dc_served, res_served). If even residential fails, the company gets status: "error" with the reason — never a fake empty result.
  • Cloudflare also judges the TLS fingerprint, and changes its mind. Which browser fingerprints Indeed serves moves over time (2026-10-05: one of the client library's profiles; 2026-11-17: 19 of 53; 2026-11-18: Firefox 133 dropped). Headers that contradict the fingerprint are refused even on a served IP, so each fingerprint goes with its own browser's headers. Every sweep spreads its sessions over three fingerprints (Firefox 154, Safari 27, Chrome 150), the first page served sets the run's, and when a whole wave is refused before any page was served the same IPs are asked again with five more from four browser families.
  • A refusal half-way keeps what was delivered. Reviews are written to the dataset page by page. If Indeed stops answering on page 40, the first 39 pages are yours and charged, and the status row says partial. The same holds when you abort a run or it reaches your maximum cost.
  • A run the platform starts again bills nothing twice. A resurrected run, or one Apify migrates to another server mid-run, starts from the top with its dataset as it was: it reads back the reviews it already delivered, skips the companies it finished and carries on with the others from the page it stood on (saved every minute), so no review is delivered or charged twice and no status row is written twice.
  • A name several companies share goes to Indeed's own pick. Meta is a small company of that name; the Facebook parent is Meta-dd1502f2 (which is where Facebook resolves). The company, company_name and review count on the free company row show what you got — pass the /cmp/ URL when it has to be exact.
  • onlyNewReviews starts from a baseline. The first run delivers the newest maxReviewsPerCompany reviews and everything older is history, not news: it is never delivered by a later run. A run also delivers at most maxReviewsPerCompany new reviews, so keep that number above what the company receives between two runs.
  • A sub-rating of 0 means "not rated", not zero stars. Indeed writes 0 when a reviewer skipped a category. Those become null here, so an average is not silently dragged to the floor.
  • Counts differ by domain. review_count is Indeed's worldwide total; found_review_count what the domain you read lists. When the countries you read hold at most half of a company's reviews, reviews_outside_countries says how many you did not get.
  • Indeed's country sites ignore a city filter. uk.indeed.com/cmp/Tesco/reviews?floc=London lists all 1,381 UK reviews of Tesco; www with fcountry=GB&floc=London lists the 28 London ones (2026-11-18). So a pasted URL's filters are always sent to www, with its country as fcountry.
  • A handful of Indeed hosts are not review sites. my.indeed.com is Indeed's account subdomain, for one; such countries (Malaysia, Korea, …) are read on www.indeed.com through Indeed's own country filter instead.
  • Dates in a language we have not met. Every domain in the list was checked, and the raw string is always kept in review_date_text; if a format ever changes, review_date becomes null rather than wrong, and sinceDate keeps such a review rather than silently dropping it.
  • Reviews are moderated by Indeed. This is its published corpus, not every review ever submitted.

FAQ

How do I find a company's Indeed slug?

You do not need it: type the company's name, or give its website. Indeed itself redirects Home Depot to The-Home-Depot, Amazon to Amazon.com and Ernst & Young to Ey, and the rows carry the slug it chose. Legal suffixes are fine too: Indeed resolves most of them (Walmart Inc, The Home Depot, Inc.), and when it does not, the actor asks again without the suffix and then under the one company of Indeed's autocomplete that has exactly that name (Amprius Technologies Inc → Amprius-Technologies,-Inc.). A merely similar name is only listed in suggestions, never read. If you want to be exact — some names are shared by several companies — open the company on Indeed and paste its URL, or just the part after /cmp/. Either way you get Indeed reviews by company rather than by keyword.

Can I get all Indeed reviews for a company, not just the US ones?

Yes: set countries to ["all"] and maxReviewsPerCompany to 0. That reads www.indeed.com with Indeed's worldwide filter and returns essentially the complete corpus in one walk. A run that reads fewer countries says so when they miss most of a company's reviews (reviews_outside_countries on the status row, and the run card).

Does it return Indeed interview questions?

Yes, from 2026-12-02, with includeInterviews: true or an /interviews URL: the questions people asked about interviewing at the company, each with its top answer, read from the Interviews topic of the company's Q&A and billed as Q&A rows. Indeed's /interviews URL now redirects to the company overview (2026-11-18), so that topic is where its interview questions are published.

How do I export Indeed reviews to CSV?

In the Console, open the run's Export dialog, pick the Reviews view and choose CSV or Excel: 19 named columns. A view picks columns, not rows, so by default the file also has a line for each company row and each status row (100 Google reviews gave 102 lines, run As5OCqbxtvCGrK4Ud). For a file of reviews only, one per line, start the run with "statusRows": "none" and "includeCompanyRow": false: 17 lines for 17 reviews (run MpaBxwVc1kha0upsL; the quick start's curl call gave the same file in 5 seconds, run xnJVWeBf87QbLG8Sn); what happened to each company is then in the run's STATUS record. From the API add &format=csv&view=reviews to the dataset items call, as in the curl example above. Without a view the CSV also flattens the company row's nested fields into hundreds of mostly empty columns: 987 for 100 Google reviews (run X6oDH8IMJdfAilihc).

For a table of employers, export the Companies view: one flat line per company, with its rating, review counts, star histogram, category averages, happiness scores, CEO approval with how many rated it, and profile: 59 columns for Google on 1.17 (run 8ox5bEXjZVhWjni1V), one more (ceo_approval_count) from 1.18. Up to 1.16 that view flattened each company's job, competitor and Q&A lists into 649 to 723 columns (run vnQzMEmDwygv8Edmu: 649); the lists stay in the JSON. JSON is not affected by any of this, and every row of a type carries the same fields, so the columns are stable.

How do I get the reviews into Google Sheets?

Three ways. With no other tool: put =IMPORTDATA("https://api.apify.com/v2/acts/kestrel~indeed-company-reviews/runs/last/dataset/items?token=<YOUR_APIFY_TOKEN>&format=csv&view=reviews&status=SUCCEEDED") in a cell, and the sheet shows the reviews of your last successful run (keep the sheet private: the formula holds your token). By hand: Export → Reviews view → CSV, then File → Import. On a schedule: Make or n8n, as in Integrations, with onlyNewReviews: true so each run appends only new reviews.

Can Claude, ChatGPT or another AI agent use it?

Yes, through Apify's MCP server: add https://mcp.apify.com?tools=kestrel/indeed-company-reviews to the agent's MCP settings (OAuth, or your Apify token as a Bearer header), or run npx -y @apify/actors-mcp-server --tools kestrel/indeed-company-reviews locally with APIFY_TOKEN set. The agent gets one tool, kestrel--indeed-company-reviews, whose parameters are this actor's inputs, and can ask in plain words ("the 50 lowest-rated Starbucks reviews from former employees"). Small calls end within the server's 30-second wait: a 5-review tool call answered in 9.5 seconds (run Gu3xbSFO4ABKoZoLJ), and 100 reviews take about 7 seconds of run time (run As5OCqbxtvCGrK4Ud). compact: true, statusRows: "problems" and includeCompanyRow: false leave the agent the reviews and only the status rows that say something went wrong (run 7HKZIHZDSSUfpN4rK: 3 rows of 16 fields). A tool call takes no maximum cost, so maxReviewsPerCompany (default 100, $0.30 a company) is what bounds what an agent spends.

Can I get only the negative reviews?

Set sort to rating_asc and maxRating to 2. The lowest-rated reviews arrive first and the walk stops at the first page above the ceiling, so you pay only for the complaints.

Can I find the employers of an industry or a place without naming them?

Yes. searchCompanies: ["trucking"] with searchLocation: "Texas" asks Indeed's own company search, the one behind its "Popular trucking companies in Texas" page, and reads the first maxCompaniesPerSearch companies that have reviews (default 10) like companies you named: the same filters and caps. The search's free status row lists all the companies Indeed showed (up to 50, with ratings and review counts) in search_results, and each company row names its search in found_by: "trucking" in Texas listed 50 (run XB1qXnZeOfwcukt6i). A listing named exactly like the keyword with fewer than 100 reviews is skipped as a namesake. For company profiles only, add includeReviews: false: every row is then free (kaix's form did exactly that, five trucking companies, nothing charged: run Yf7COjFjwLPH1DpKg).

Can I get only the reviews about management, or in one language?

Yes, with Indeed's own filters: topic: "management" (or work_life_balance, compensation_benefits, job_security_advancement, culture_values) keeps the reviews Indeed files under that topic, and language: "es" the reviews written in Spanish. Indeed applies both, so the run reads only those pages; a topic's status row carries the topic's own count (Google, US, management: 994 of its 4,068 US reviews on 2026-11-19, run RPaY1jAHPbMN10MaX), while a language has none (run IAJUHwgdW5cCXfrXL).

Does it return the employer's replies?

Yes — employer_reply and employer_reply_date, when the company has responded. Indeed employer replies are missing from most reviews, which is itself a useful metric.

How do I monitor new reviews every day?

Schedule the actor with onlyNewReviews: true. The first run is the baseline; every later run stops at the first page it already knows and delivers only the reviews written since, so a quiet day costs nothing. Add maxRating: 2 to be told about new complaints only.

My run stopped early (time limit, maximum cost, abort). Can I continue it?

Yes. A run stops starting pages a little before its time limit, at its maximum cost, after a graceful abort, or when Indeed refuses a company; it keeps and bills what it delivered, and — when companies are left — its run card names the id to continue from and, at the maximum cost, how far back each cut company got. Start the same input with "resumeFromRun": "<that run id>" (and a higher maximum cost if that was the stop): every company carries on where it stopped, under the same per-company cap, and nothing that run delivered is delivered or charged again. Companies it finished are listed with status done. A run aborted at once, or one Apify migrated, is best resurrected (the run's page → Resurrect): it reads back what it delivered and carries on, billing no review twice (run JtMesmY9uv4fWaqRW: aborted at once at 180 reviews, resurrected to 300 unique, 300 charged).

Is there an Indeed employee reviews API?

Not a public one. This actor is the practical substitute: Indeed reviews without an API key, a stable JSON contract over the public company pages, paging and retries handled for you.

What happens if a company has no reviews?

You get