🕵️ Contact Finder SDK

A robust, high-concurrency async Python SDK for the Website Contact Finder API. Effortlessly scrape emails, phone numbers, social media links, and rich company metadata from thousands of websites.

✨ Key Features

• 🚀 Async & High Concurrency: Processes hundreds of URLs simultaneously using asyncio and aiohttp.
• 🛡️ Crash-Proof Resumability: Caches raw API responses locally. If your script crashes, it resumes exactly where it left off without wasting your RapidAPI quota.
• 🔄 Self-Healing Storage: Automatically detects missing files and rebuilds them. If you delete a CSV, it rebuilds it from the JSONL. If you delete the JSONL, it rebuilds it from the CSV.
• 📊 Dynamic CSV Headers: Uses atomic file rewrites to dynamically add new columns if the API returns unexpected fields, without corrupting existing data.
• 🗂️ Multi-Project Isolation: Run multiple scrapers at once. Each project_name gets its own isolated data folder.
• 🧠 Smart Callback Routing: Triggers your on_success callbacks instantly for cached URLs without making redundant API calls.

---

📦 Installation

pip install contact-finder-sdk

🔑 Prerequisites & API Key Setup

This SDK connects to the Website Contact Finder API hosted on RapidAPI. You will need an API key to start scraping.

1. Subscribe to the Free Plan Visit our RapidAPI Pricing Page.

• Click "Subscribe" or "Start Free Plan".
• Log in to RapidAPI (using "Continue with Google" is the fastest).
• Select the Free Plan (usually the first one on the left) and click "Subscribe" in the top right corner. (No credit card required!)

2. Copy Your API Key Once subscribed, go to the Endpoints tab on the API dashboard. Click on any endpoint, and look at the code snippet on the right side of the screen. Copy the long string value for x-rapidapi-key.

💡 Need help? If you get stuck or have questions about the API, message me directly on WhatsApp or Telegram and I'll walk you through it personally!

---

🚀 Quick Start (Zero Configuration Required)

We designed this SDK to get out of your way. You don't need to write loops, handle callbacks, or manually parse and save files.

Just pass your URLs, and the SDK automatically scrapes, extracts emails/phones/socials, and drops perfectly formatted files into a folder named after your project.

from contact_finder_sdk import ContactFinder

target_urls = ["https://rapidapi.com", "https://github.com"]

# 1. Initialize with your credentials and project name
finder = ContactFinder(api_key="YOUR_RAPIDAPI_KEY", concurrency=5, project_name="my_first_scrape")

# 2. Pass the URLs. The SDK handles everything else automatically.
finder.run(urls=target_urls)

---

📁 Understanding Your Output Files

As soon as the Quick Start finishes, you will notice a new folder in your directory: ./data/my_first_scrape/.

Because auto_store=True by default, the SDK automatically generates the following beautifully structured files:

File	Description
full.jsonl	The Master Database. Contains all parsed, flattened records.
raw.jsonl	The Cache. Raw API responses used to prevent re-scraping.
full.csv	A spreadsheet version of the master database.
contact.csv	A lightweight spreadsheet with only core contact info (Emails, Phones, Socials).
failed.log	A log of URLs that failed to scrape, with error reasons and timestamps.

🛡️ Data Safety & Resumability

You don't need to worry about data corruption or crashing scripts:

1. Auto-Rebuilding: If you accidentally delete full.csv, the SDK detects it on the next run and automatically rebuilds it from full.jsonl. If you delete the JSONL, it rebuilds it from the CSV.
2. Dynamic Headers: If the API returns a brand new field that wasn't in your original CSV header, the SDK uses atomic file rewrites to dynamically add the new column without corrupting your existing rows.
3. Resumability: If your script crashes halfway through, simply run it again. It reads raw.jsonl, skips the URLs it already finished, and picks up exactly where it left off.

---

📂 Reading from a CSV or TXT File

If you have a massive list of URLs, you don't need to load them into a Python list first. Just pass a .csv or .txt file. The SDK will automatically read the file, extract the URLs, and process them.

finder.run(
    file_path="my_urls.csv",    # Path to your CSV or TXT file
    url_field_name="url",       # The name of the column containing the URLs
)

---

🗂️ Multi-Project Isolation

One of the most powerful features is Project Isolation. If you are managing multiple clients or campaigns, simply assign a different project_name to each instance.

The SDK will automatically create separate, isolated folders for each project inside a local ./data/ directory. Their caches, CSVs, and logs will never mix.

# Instance 1: Client A
finder_a = ContactFinder(api_key="YOUR_KEY", project_name="client_A")
finder_a.run(urls=["https://client-a.com"])
# Automatically saves to: ./data/client_A/

# Instance 2: Client B
finder_b = ContactFinder(api_key="YOUR_KEY", project_name="client_B")
finder_b.run(urls=["https://client-b.com"])
# Automatically saves to: ./data/client_B/

---

🪝 Real-time Callbacks & Data Parsing

Want to save data directly to your database as it arrives? Pass an on_success callback.

Smart Routing: If the SDK finds a URL in its local cache, it skips the API call (saving your RapidAPI quota) but still instantly triggers your callback with the cached data!

def save_to_my_db(data):
    print(f"✅ Scraped {data.get('domain')}")
    print(f"📧 Emails: {data.get('emails')}")

finder.run(
    urls=["https://rapidapi.com"],
    parse_mode="contact",       # Controls what data is passed to your callback
    on_success=save_to_my_db    # Triggered for BOTH new and cached URLs
)

🔍 Parse Modes Explained

• "contact": Returns a flattened dictionary containing only core contact info (Emails, Phones, Social Media links). Highly recommended for on_success callbacks.
• "full": Returns a massive flattened dictionary containing Contact info + Company Profile + Visuals + Meta tags.
• "raw" (Default): Returns the exact, untouched JSON dictionary from the RapidAPI response.

---

⚡ Advanced: Streaming Data & Payload Optimization

By default, auto_store=True and all API extraction flags (profile, visuals, meta) are set to "true". This fetches everything and safely writes it to local CSV/JSONL files, which is the recommended approach for most users.

However, if you are building a high-volume pipeline and only need to stream data directly into your database via on_success, you can optimize your workflow in two ways:

💡 Pro-Tip for High-Volume Pipelines:

1. Speed up your local workflow: Set auto_store=False. Bypassing local file generation eliminates disk I/O, significantly speeding up your Python script's execution when you only need the on_success callback.
2. Reduce payload size & save bandwidth: Change the extraction flags (profile, visuals, meta) to "false". This tells the API to skip heavy extraction, drastically reducing the RapidAPI response payload size and saving massive amounts of bandwidth when processing millions of requests. (Note: This reduces network transfer size, but does not significantly speed up the API's internal scraping time).

finder.run(
    urls=target_urls,
    on_success=save_to_my_db,   # Process data in real-time
    auto_store=True,            # Default: True. Set to False to skip disk I/O and speed up execution.
    parse_mode="contact",       # Default: "raw". Use "contact" for emails/phones.
    
    # API Overrides (Defaults are all "true" to fetch complete data for auto_store)
    # Change these to "false" to reduce response payload size and save bandwidth:
    profile="true",             # Default: "true". Set to "false" to skip company profile extraction.
    visuals="true",             # Default: "true". Set to "false" to skip image/logo extraction.
    meta="true"                 # Default: "true". Set to "false" to skip meta tag extraction.
)

---

⚙️ Complete Configuration Reference

The Quick Start uses smart defaults that work perfectly for 95% of use cases. But if you want granular control, finder.run() accepts the following parameters:

Parameter	Type	Default	Description
urls	list[str]	None	A direct list of URLs to scrape.
file_path	str	None	Path to a .csv or .txt file containing URLs.
url_field_name	str	"url"	The column name to look for if using a CSV file.
dedup_key	str	"url"	The key used to check for duplicates.
parse_mode	str	"raw"	How to format the callback data ("raw", "contact", or "full").
on_success	Callable	None	A function to execute immediately when a URL succeeds.
auto_store	bool	True	If True, automatically saves outputs to local files.
**api_kwargs	kwargs	{}	Override default RapidAPI params (e.g., profile="false").

---

📄 License

This project is licensed under the MIT License - see the LICENSE file for details.