Glassdoor Analyzer

Glassdoor Analyzer

Scrape Glassdoor for company reviews, salary data, interview experiences, and culture signals. Glassdoor has aggressive bot detection — always run via the cloud browser.

Output directory: ~/Desktop/{slug}_glassdoor_{YYYY-MM-DD}/

What It Extracts

Company Overview

• Overall rating (out of 5)
• Rating breakdown: Culture, Work/Life Balance, Management, Compensation, Career Growth
• CEO approval rating
• Recommend to a friend %
• Recent rating trend (improving/declining)

Reviews

• Top positive reviews (what employees praise)
• Top critical reviews (pain points, churn reasons)
• Most mentioned pros and cons
• Management-specific complaints

Salary Data

• Roles with salary ranges
• Compensation by seniority level
• Location-based salary variations

Interview Data

• Difficulty rating
• Common interview questions
• Process length and stages
• Offer rate

Pipeline

Run each step as a separate goal-driven task. Capture the watch-task id from BROWSER_TOOL_CREATE_TASK and poll BROWSER_TOOL_WATCH_TASK for the output. If a task hits a CAPTCHA or login wall, stop the task and split the goal smaller.

Step 1: Find the company URL

composio execute COMPOSIO_SEARCH_WEB -d '{ "query": "site:glassdoor.com {COMPANY_NAME} reviews" }'
# Pick the URL of the form: glassdoor.com/Reviews/{company}-Reviews-{ID}.htm

Step 2: Reviews — overall + recent

composio execute BROWSER_TOOL_CREATE_TASK -d '{
  "task":     "Capture: overall rating; the five sub-ratings (Culture & Values, Work/Life Balance, Senior Management, Compensation & Benefits, Career Opportunities); CEO approval %; Recommend-to-friend %. Then sort reviews by Most Recent and capture the latest 30 reviews — each as { title, rating, pros, cons, role, date }. Stop when 30 are captured or the page stops loading more.",
  "startUrl": "https://www.glassdoor.com/Reviews/{COMPANY_SLUG}-Reviews-{ID}.htm"
}'

Step 3: Salaries

composio execute BROWSER_TOOL_CREATE_TASK -d '{
  "task":     "Extract every visible role with its level (junior/mid/senior/staff if shown), low/median/high salary, and total comp if shown. Return a JSON array.",
  "startUrl": "https://www.glassdoor.com/Salary/{COMPANY_SLUG}-Salaries-{ID}.htm"
}'

Step 4: Interviews

composio execute BROWSER_TOOL_CREATE_TASK -d '{
  "task":     "Capture: difficulty rating, offer rate, average process length, and the 15 most upvoted interview questions with the role they apply to.",
  "startUrl": "https://www.glassdoor.com/Interview/{COMPANY_SLUG}-Interview-Questions-{ID}.htm"
}'

Output Format

---
company: {Name}
glassdoor_url: {URL}
overall_rating: {N}/5
total_reviews: {N}
researched_at: {ISO date}
---

## Ratings Breakdown
| Category | Score |
|----------|-------|
| Culture & Values | {N}/5 |
| Work/Life Balance | {N}/5 |
| Senior Management | {N}/5 |
| Compensation & Benefits | {N}/5 |
| Career Opportunities | {N}/5 |

**CEO Approval**: {N}%
**Recommend to Friend**: {N}%

## Top Pros (What Employees Love)
- "{Pro quote}" — {role}, {date}
- "{Pro quote}" — {role}, {date}

## Top Cons (Pain Points)
- "{Con quote}" — {role}, {date}
- "{Con quote}" — {role}, {date}

## Salary Ranges
| Role | Level | Salary Range |
|------|-------|-------------|
| Software Engineer | Senior | ${min}K – ${max}K |
| Product Manager | Mid | ${min}K – ${max}K |

## Interview Insights
- **Difficulty**: {Easy/Medium/Hard} ({N}% rating)
- **Process**: {description of stages}
- **Common Questions**:
  - "{question}"
  - "{question}"

## Culture Signals
| Signal | Finding |
|--------|---------|
| Management quality | {assessment} |
| Engineering culture | {assessment} |
| Remote/hybrid | {policy} |
| Growth opportunities | {assessment} |

Key Use Cases

1. Job research — understand a company's culture before applying
2. Sales intelligence — find pain points to personalize outreach ("we saw your engineers rate management 2.1/5...")
3. Competitive talent analysis — is your competitor's team happy or about to leave?
4. Hiring research — understand what candidates care about at competitors
5. Culture benchmarking — compare ratings across companies in a space