🏢 Firmographic Enrichment Actor

Automatically enrich company data from website URLs using Google SERP AI Mode + Groq LLM.

Input a single URL → Get structured firmographic data:

Field	Example
employee_min / employee_max	500 / 1000
revenue_min_usd / revenue_max_usd	10000000 / 50000000
funding_stage_include	"series-b"
founded_year_min / founded_year_max	2010 / 2010
company_status	"active"
total_funding_usd	150000000
industry	"Fintech"
headquarters	"San Francisco, CA"
notable_investors	"Sequoia, Andreessen Horowitz"
data_confidence	"high" / "medium" / "low"

---

🚀 Quick Start

1. Get Free API Keys

SerpApi (Google SERP):

• Sign up at https://serpapi.com
• Free tier: 100 searches/month
• Used for AI Overview + organic results

Groq (LLM):

• Sign up at https://console.groq.com
• Free tier: generous rate limits
• Model: llama-3.1-8b-instant (fastest, free)

2. Configure Input

Set these environment variables in your Actor settings:

• SERP_API_KEY
• GROQ_API_KEY

Then provide the input:

{
  "websiteUrl": "https://stripe.com"
}

---

📁 File Structure

firmographic-actor/
├── .actor/
│   ├── actor.json           # Actor metadata
│   └── input_schema.json    # Input field definitions + UI
├── src/
│   ├── main.js              # Actor entrypoint, orchestration
│   ├── serpService.js       # Google SERP API queries
│   ├── groqService.js       # Groq LLM extraction
│   └── csvUtils.js          # CSV output helpers
├── storage/
│   └── key_value_stores/
│       └── default/
│           └── INPUT.json   # Local dev input
├── test_local.js            # Local test runner
├── Dockerfile               # Apify deployment container
├── package.json
└── README.md

---

🔄 How It Works

URL Input
    │
    ▼
SerpApi (3 queries per company)
    ├── "company.com employees revenue funding"
    ├── "company.com founded year funding stage raised"  
    └── "company.com headquarters industry sector"
    │
    ▼
Text aggregation
    ├── AI Overview blocks
    ├── Knowledge Graph attributes
    └── Organic result snippets
    │
    ▼
Groq LLM (llama-3.1-8b-instant)
    └── Structured JSON extraction
    │
    ▼
Output
    ├── Apify Dataset (streaming, per record)
    └── CSV file (Key-Value Store)

---

💻 Local Development

# Install dependencies
npm install

# Set env vars
export SERP_API_KEY=your_serpapi_key
export GROQ_API_KEY=your_groq_key

# Run local test (3 example companies)
node test_local.js

# Or run as full Apify actor locally
npx apify-cli run

---

📤 Output

Apify Dataset

Every record is pushed to the dataset in real-time. Download as JSON, CSV, or XLSX from the Apify console.

Key-Value Store CSV

A firmographic_results.csv file is saved to the actor's Key-Value Store. Download URL format:

https://api.apify.com/v2/key-value-stores/{STORE_ID}/records/firmographic_results.csv

---

⚡ Performance Tips

• Rate limits: SerpApi free = 100 searches/month (3 per company = ~33 companies free)
• Groq: Very generous free tier, effectively unlimited for this use case
• Concurrency: Keep at 2-3 to avoid hitting SerpApi rate limits
• Delay: 1000ms between requests is safe for free tiers

🆓 Free Alternative APIs

Purpose	Service	Free Tier
Web search	SerpApi	100/month
Web search (alt)	ValueSERP	100/month
LLM extraction	Groq	Very generous
LLM extraction (alt)	Together AI	$25 free credits
Company data	Clearbit (now HubSpot)	Limited free

---

🛠 Deploying to Apify

# Install Apify CLI
npm install -g apify-cli

# Login
apify login

# Push to Apify
apify push

# Run on Apify cloud
apify call YOUR_USERNAME/firmographic-enrichment