This repository will be replaced with a new implementation in the near future:
go to https://github.com/noscrape/noscrape-node for more informations

Concept

The primary mechanism behind noscrape is the utilization of any true-type font. From this, noscrape generates a new version with shuffled unicodes, ensuring that it's impossible to reverse-calculate them. This means that both strings and integers are obfuscated and can only be deciphered using the generated obfuscation-font.

While the glyph-paths inside the font cannot be entirely removed, they are obfuscated by randomly shifting them slightly. This makes it challenging to reverse-calculate them, but it's not entirely impossible, especially for machine learning algorithms. The developers are open to suggestions for improving this aspect.

Use-Cases

In an era where artificial intelligence is becoming increasingly integral to our daily lives, it's important to remember that AI thrives on data, and your data is a valuable commodity that shouldn't be given away lightly.

1. Anti-Scraping Measures for Websites:
   Implement noscrape on your website to protect against web scrapers. This can be particularly useful for content that is unique to your site, so you wish to prevent it from being copied or used without permission.
   
   • sport results
   • betting results
   • prices (e-commerce)
   • geo-information
   • ...
     
     
2. Protecting Sensitive Data
   Use noscrape to obfuscate sensitive information such as personal identifiers, financial details, or confidential text in a way that is visually accessible but protected against scraping and automated data extraction tools.
   
   
3. Reduce Bot interactions
   Once your data is protected by noscrape it makes no sense to scrape them and one can reduce the number of bot interactions and so to lower costs at the end.
   
   
4. Secure Applications
   In applications where data security is paramount, such as in banking or healthcare apps, noscrape can be used to display information in a secure manner.
   • PIN/TAN numbers
   • Bot protection (captcha)

Example code

Live demo

Installation

To install the @noscrape/noscrape package, simply run the following command in your project directory:

npm install @noscrape/noscrape

Basic Usage

Server-side

const { Noscrape } = require('@noscrape/noscrape');

// create noscrape instance
const noscrape = new Noscrape('path/to/font.ttf', { options })

// obfuscate data
const number = noscrape.obfuscate(123);
const string = noscrape.obfuscate("noscrape");
const object = noscrape.obfuscate({ title: "noscrape", text: "obfuscation" });

// generate obfuscation font buffer after all obfuscation is done
const font = noscrape.getFont();

then provide font and data to the client/frontend

Client-side

<style> 
    @font-face {        
        font-family: 'noscrape-obfuscated';        
        src: url('data:font/truetype;charset=utf-8;base64,${font.toString("base64")}');    
    }
</style>

The font is delivered in a buffer format. To utilize it in our web pages, we convert it into a base64 URL and embed it within a custom @font-face declaration. Once this is done, we can display the obfuscated data using the specified font-family in our styles.

<span style="font-family: noscrape-obfuscated">
    <div>{ object.title }</div>
    <div>{ object.text }</div>
</span>

IMPORTANT NOTE

Bots might not be able to process obfuscated text, which can lead to unpredictable analytics results. Therefore, it's advised not to use this technology on content that's essential for indexed pages. The obfuscation process takes some time (around 50-60ms on standard machines). For API requests, it's recommended to put the obfuscation logic into a scheduled task and reuse the results, rather than recalculating everything for every request.

Options

• Strength (obfuscation strength multiplier)
  Default is 1. Values below 0.1 are not recommended as paths can be easily reverse-calculated. Values over 10 might not look visually appealing.
  

• Character Range
  This defines the character range used for encryption. Options include:

  • PRIVATE_USE_AREA (default)
  • LATIN
  • GREEK
  • CYRILLIC
  • HIRAGANA
  • KATAKANA
    
    

• Low Memory
  This option is for situations with limited memory where noscrape cannot load the provided font file. Default is false.

Contributions

The developers welcome contributions, issues, and feature requests. If you've used this package and fixed a bug, they encourage you to submit a PR.

License

The package is licensed under the MIT License by Bernhard Schönberger.