Skip to content

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

36 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

nastaliq-captcha

Generate CAPTCHA images using random numbers spelled out in Persian (Farsi) words, rendered in the Nastaliq script with randomized colors, noise, and rotation to resist OCR.

Note: This README assumes the package is published as nastaliq-captcha. Adjust the install command below if your package.json uses a different name.


Features

  • 🔢 Random number generated with crypto.randomInt (cryptographically secure, not predictable)
  • 🖋 Rendered as Persian words in the Nastaliq font, not digits — harder for digit-based OCR to crack
  • 🎨 Randomized background/foreground colors with a built-in contrast check, so the text never becomes unreadable
  • 🌪 Visual noise: scattered dots + interference lines to defeat naive image-based bots
  • 📐 Auto-fitting font size and slight random rotation for additional distortion
  • 🖼 Outputs a plain PNG Buffer — works with any web framework

Installation

npm install nastaliq-captcha

This package depends on canvas for image rendering, which uses native bindings. On most systems this installs automatically via npm install, but if you hit build errors, see node-canvas's platform-specific setup guide (Cairo/Pango system libraries are required on Linux/macOS).


Quick start

import { NastaliqCaptcha } from "nastaliq-captcha";

const captcha = new NastaliqCaptcha();

const { image, answer } = captcha.generate();
// image  -> PNG Buffer, send this to the client
// answer -> Number, e.g. 419

Custom dimensions

const { image, answer } = captcha.generate({
  width: 240,
  height: 80,
  difficulty: 5,
});

Both width and height are optional and independently default to 200 and 60 respectively.


⚠️ Important: handling the answer securely

generate() returns both the rendered image and the number answer in the same object. This is intentional — the package itself doesn't assume how you store sessions — but it means the calling code is responsible for keeping answer away from the client.

If you send the full { image, answer } object straight back in an API response, the CAPTCHA provides no protection at all: anyone can read the answer directly out of the response body without ever looking at the image.

✅ Correct usage pattern

Generate the challenge, store the answer server-side (session, cache, or database) keyed by some challenge ID, and send only the image to the client:

import { NastaliqCaptcha } from "nastaliq-captcha";
import { randomUUID } from "crypto";

const captcha = new NastaliqCaptcha();

// --- On challenge creation (e.g. GET /captcha) ---
app.get("/captcha", (req, res) => {
  const { image, answer } = captcha.generate();
  const challengeId = randomUUID();

  // Store server-side only — e.g. in Redis, a session store, or a short-lived DB row.
  // Example using an in-memory session for illustration:
  req.session.captcha = { id: challengeId, answer };

  res.set("Content-Type", "image/png");
  res.json({
    challengeId,
    image: image.toString("base64"), // or stream the Buffer directly
  });
});

// --- On verification (e.g. POST /captcha/verify) ---
app.post("/captcha/verify", (req, res) => {
  const { challengeId, userAnswer } = req.body;
  const stored = req.session.captcha;

  if (!stored || stored.id !== challengeId) {
    return res.status(400).json({ valid: false, reason: "Challenge expired or not found" });
  }

  const valid = stored.answer === userAnswer;
  delete req.session.captcha; // one-time use

  res.json({ valid });
});

❌ Avoid this

// DON'T: ships the answer to the client alongside the image.
app.get("/captcha", (req, res) => {
  const { image, answer } = captcha.generate();
  res.json({ image: image.toString("base64"), answer }); // anyone can read `answer` here
});

API

new NastaliqCaptcha()

Creates a new CAPTCHA generator instance. No configuration needed at construction time.

.generate(config?)

Generates a new random challenge and renders it as an image.

Param Type Default Description
config.width number 200 Output image width in pixels
config.height number 60 Output image height in pixels
config.difficulty number 5 Visual distortion level from 0 (minimal noise) to 10 (maximum noise). Controls the density of interference dots and lines, and the degree of text rotation.

Returns { image: Buffer; answer: string }

Field Type Description
image Buffer PNG image buffer
answer number The correct answer (e.g. 23). Handle this server-side only — see security section above.

How it works

  1. CaptchaCore picks a random number and converts it to Persian words via numberToPersian().
  2. CanvasRenderer draws that text onto a canvas with a randomized, contrast-checked color pair, adds noise (dots + curved lines), fits the font size to the canvas, and applies a slight random rotation before exporting a PNG buffer.
  3. NastaliqCaptcha ties the two together and exposes the single generate() entry point.
NastaliqCaptcha.generate()
  ├─ CaptchaCore.create()        → { answer, raw }
  └─ CanvasRenderer.render()     → PNG Buffer

Testing

npm install -D jest ts-jest @types/jest typescript @types/node
npx jest

Tests mock the native canvas dependency, so the suite runs without needing Cairo/Pango or the Nastaliq font file installed.


Requirements

  • Node.js (with crypto.randomInt support — Node 14.10+)
  • A Nastaliq-compatible .ttf font file at src/assets/fonts/nastaliq.ttf (bundled with the package)

License

Nastaliq CAPTCHA is source-available package. The source code is publicly visible for transparency and contribution purposes, but redistribution, derivative works, and competing products are prohibited without permission.

About

CAPTCHA image generator rendering numbers as Persian words in Nastaliq script, with randomized colors, noise, and rotation to resist OCR.

Topics

Resources

Stars

Watchers

Forks

Releases

Used by

Contributors

Languages