Skip to content

README.md

Latest commit

 

History

History
66 lines (49 loc) · 1.97 KB

README.md

File metadata and controls

66 lines (49 loc) · 1.97 KB

Client-Side Human vs Bot Detection

Overview

This repository contains a client-side script that detects whether the current browser session is likely automated or human-driven.
It runs entirely in the browser and does not rely on third-party bot detection services.

Live Demo: https://cside-assesment.vercel.app/

Detections Implemented

Static Detections

  • navigator.webdriver check
  • User-Agent inspection for headless/automation hints
  • Browser plugin anomalies
  • Language settings anomalies
  • WebGL vendor/renderer blacklist
  • Touch support vs User-Agent mismatch
  • Timezone availability
  • Screen resolution sanity check
  • Device specs (CPU cores, memory)
  • Chrome app detection
  • Permissions API presence
  • WebRTC availability

Behavioral Signals

  • Focus/blur patterns
  • Window resize events
  • Touch gestures

Activity Probe

  • Monitors mouse movements, clicks, keyboard inputs, and scroll events for 5 seconds
  • Analyzes timing variance to detect overly regular behavior

Summary and Risk Levels

Signals are aggregated into a risk score and mapped to status levels:

Level Suspicion
Human Low
Suspicious Medium
Automated High

Expected False Positives

  • Users with restricted browsers or strict privacy settings
  • Users with unusual interaction patterns (very slow or very fast inputs)
  • Mobile or touch devices with unconventional configurations

Repository Contents

  • app/page.tsx – Demo page with UI and status badge
  • app/lib/botDetector.ts – Detection script implementing all signals
  • README.md – Project overview and instructions

Usage

Simply open the demo page in a browser and the script will automatically evaluate whether the session is human or automated.

Results are displayed on the status badge and can also be inspected via the browser console.