Scavio Glassdoor

Glassdoor via Scavio

Resolve a company name to its Glassdoor employer id, then read the employer profile, employee reviews and salary percentiles by job title. All endpoints return structured JSON.

When to trigger

Use this skill when the user asks to:

• Look up a company's Glassdoor rating, star distribution, CEO approval and per-category scores
• Read employee reviews with pros, cons, advice and the employer's response
• Pull salary percentiles (P10-P90, base and total pay) by job title for an employer
• Benchmark compensation or employer brand across companies
• Turn a company NAME into the employer id Glassdoor's URLs use

Note: Glassdoor is the slowest surface on this API. A profile call runs 3-47 seconds, reviews around 75, salaries around 41, and a call that ultimately fails can take up to ~170 seconds before it reports a 502. Set a client timeout of at least 180 seconds and expect the occasional retry.

Setup

Get a free API key at scavio.dev (50 free credits to get started, no card required):

export SCAVIO_API_KEY=sk_live_your_key

Every request is a POST with a JSON body and:

Authorization: Bearer $SCAVIO_API_KEY

Endpoints

Base URL: https://api.scavio.dev. Every Glassdoor endpoint costs 1 credit.

Workflow

Start at /companies. /company, /reviews and /salaries all need an employer_id that exists only inside Glassdoor's /Overview/ URLs.

1. Resolve the name: call /glassdoor/companies with query. Each result row carries employer_id, name and url.
2. Profile: call /glassdoor/company with that employer_id. Keep the reviews_url and salaries_url it returns.
3. Reviews and salaries: call /glassdoor/reviews and /glassdoor/salaries passing those saved URLs back as url. That is the single-fetch path - addressing them by employer_id instead makes the API resolve the case-sensitive /Reviews/ and /Salary/ slugs off the profile first, which is two upstream fetches for the same one credit and roughly twice the latency on the slowest surface here.

employer_id must be a string - a JSON number is rejected. 1699, E1699 and IE1699 are all accepted.

company is cosmetic. The profile resolves on employer_id alone, company is ignored entirely when url is set, and it does not satisfy the required-identifier rule: a request carrying only company is rejected.

Pagination

• /reviews has no page parameter and never will. Glassdoor's login wall caps the response at three reviews. Move the window with category and employment_status, and read filtered_review_count to see how many reviews match the current filter (against total_review_count for the employer overall). Do not build a paging loop here and never claim you can retrieve all reviews.
• /salaries pages with a 1-based page at 10 job titles per page; page_count on the response says how many pages exist.
• /companies and /company return a single response and take no paging parameter.

Parameters

Companies lookup (/companies)

Company (/company)

employer_id or url is required.

Reviews (/reviews)

employer_id or url is required. There is no page parameter. Both filters are closed sets: Glassdoor ignores an unknown value and serves the unfiltered set under a 200, so an invented filter is a credit spent on nothing. A freelance status is absent because it was never confirmed to change the result set.

Salaries (/salaries)

Examples

import requests

BASE = "https://api.scavio.dev"
# Your key from https://scavio.dev. Load it from your environment or secret
# store in real code - keep it out of source control.
API_KEY = "sk_your_key_here"
HEADERS = {"Authorization": f"Bearer {API_KEY}"}
TIMEOUT = 180          # Glassdoor is slow - do not use a short client timeout

# 1. ALWAYS START HERE - a name is not an employer id
hits = requests.post(f"{BASE}/api/v1/glassdoor/companies", headers=HEADERS,
    json={"query": "Anthropic"}, timeout=TIMEOUT).json()

employer_id = hits["data"]["results"][0]["employer_id"]   # a STRING

# 2. Profile - keep the two URLs it hands back
company = requests.post(f"{BASE}/api/v1/glassdoor/company", headers=HEADERS,
    json={"employer_id": employer_id}, timeout=TIMEOUT).json()["data"]

reviews_url, salaries_url = company["reviews_url"], company["salaries_url"]

# 3. Reviews via the saved URL - one upstream fetch instead of two.
#    Up to THREE reviews. There is no page param; move the window with filters.
reviews = requests.post(f"{BASE}/api/v1/glassdoor/reviews", headers=HEADERS,
    json={"url": reviews_url, "category": "work_life_balance",
          "employment_status": "full_time"}, timeout=TIMEOUT).json()["data"]

print(reviews["count"], reviews["filtered_review_count"], reviews["total_review_count"])

# 4. Salary percentiles by job title, 10 per page
salaries = requests.post(f"{BASE}/api/v1/glassdoor/salaries", headers=HEADERS,
    json={"url": salaries_url, "page": 1}, timeout=TIMEOUT).json()["data"]

print(salaries["page"], "of", salaries["page_count"], salaries["pay_period"])

Response shapes

Every response uses the envelope { data, response_time, credits_used, credits_remaining }. Key data fields:

• companies — count, results[] (pos, employer_id, name, url).
• company — employer_id, name, short_name, url, reviews_url, salaries_url, website, description, mission, logo, cover_photo, headquarters, size, type, revenue, stock_symbol, year_founded, industry, sector, office_count, is_active, rating, rating_max, rating_count, review_count, salary_count, ratings, industry_ratings, rating_distribution, ceo, awards, faq, breadcrumbs, reviews[] (the five server-rendered ones).
• reviews — employer_id, name, url, reviews_url, salaries_url, rating, total_review_count, rating_count, filtered_review_count, ratings, rating_distribution, highlights[] (term, type, text), reviews_by_job_title, count (at most 3), reviews[] (pos, review_id, url, title, pros, cons, advice, rating, rating_career_opportunities, rating_compensation_and_benefits, rating_culture_and_values, rating_diversity_and_inclusion, rating_senior_management, rating_work_life_balance, business_outlook, ceo_opinion, recommends, job_title, location, employment_status, is_current_employee, years_employed, published_date, helpful_count, employer_response, employer_response_date).
• salaries — employer_id, name, url, location, pay_period, currency, salary_count, job_title_count, page, page_count, count, salaries[] (pos, job_title, job_title_id, salary_count, currency, base_pay, total_pay, median_base_pay, median_total_pay, most_recent).

{
  "data": {
    "employer_id": "1699",
    "name": "Example Corp",
    "rating": 3.9,
    "total_review_count": 17798,
    "filtered_review_count": 13251,
    "count": 3,
    "reviews": [
      {
        "pos": 1,
        "review_id": "98765432",
        "title": "Good pay, heavy workload",
        "pros": "Compensation is above market and the tooling is excellent.",
        "cons": "On-call rotation is punishing during launches.",
        "rating": 4,
        "rating_work_life_balance": 3,
        "job_title": "Software Engineer",
        "location": "Austin, TX",
        "employment_status": "full_time",
        "is_current_employee": true,
        "published_date": "2026-06-14"
      }
    ]
  },
  "credits_used": 1,
  "credits_remaining": 999
}

Guardrails

• Every call is 1 credit, including one that comes back empty or times out into a 502.
• Reviews are capped at three per response. Never promise, imply or attempt a full review dump, and never invent a page parameter for /reviews. Report the cap, and use filtered_review_count / total_review_count to give the user the scale of what exists.
• Three reviews are not a sample you can generalise from. Use the aggregate blocks - rating_distribution, ratings, highlights, reviews_by_job_title - for anything sentiment-shaped, and quote the individual reviews as anecdotes.
• Chain reviews_url / salaries_url from /company whenever you already have them. Same price, half the upstream work, materially faster on a slow surface.
• employer_id must be a string. Send "1699", not 1699.
• Never send company on its own expecting it to address an employer - it is cosmetic and the request is rejected without employer_id or url.
• Salary figures are Glassdoor's estimates for the job title, not individual reported salaries. Say so whenever you quote one, along with pay_period and currency.
• Never fabricate employer ids, ratings, salary figures or review text. Only return what the API returned.
• Reviews are written by identifiable employees. Summarise them; do not try to identify who wrote one.

Failure handling

• 400 means an invalid or missing parameter - neither employer_id nor url, a numeric employer_id, a non-glassdoor.com url, or a category / employment_status outside its set. Fix and retry.
• 401 means the API key is invalid or missing. Check SCAVIO_API_KEY.
• 404 means the employer id or URL does not resolve. Re-run /companies rather than guessing another id.
• 429 means rate or usage limit exceeded. Wait before retrying. See rate limits.
• 502 / 503 mean upstream is temporarily unavailable. This is the common failure here - Glassdoor sits behind a render-only, degraded pool. A failing call can take ~170 seconds before it reports. Retry once or twice with a short pause; do not hammer it, and do not shorten the client timeout to "fail faster" - you will abandon calls that were about to succeed.
• If /companies returns nothing useful, try the legal entity name or a shorter form of the brand.
• If SCAVIO_API_KEY is not set, prompt the user to export it before continuing.

SDKs

pip install scavio==0.15.0

from scavio import ScavioClient

client = ScavioClient()  # reads SCAVIO_API_KEY

hits = client.glassdoor.companies("Anthropic")
employer_id = hits["data"]["results"][0]["employer_id"]

company = client.glassdoor.company(employer_id=employer_id)["data"]
reviews = client.glassdoor.reviews(url=company["reviews_url"], category="compensation")
salaries = client.glassdoor.salaries(url=company["salaries_url"], page=1)

npm install scavio@0.15.0

import { Scavio } from "scavio";

const client = new Scavio(); // reads SCAVIO_API_KEY
const hits = await client.glassdoor.companies({ query: "Anthropic" });
const company = await client.glassdoor.company({ employer_id: hits.data.results[0].employer_id });
const salaries = await client.glassdoor.salaries({ url: company.data.salaries_url, page: 1 });