Manhwa Agent Bot

Full Stack Developer

Impact

  • Architected 3-tier containerized system: Playwright/FastAPI scraper + n8n orchestrator + Flet desktop GUI
  • Implemented anti-bot bypass with headless Chromium, User-Agent spoofing, and 4-second rate-limit throttling
  • Designed 12-node n8n automation workflow with dual-trigger (6-hour schedule + on-demand webhook)
  • Engineered base64 image pipeline to bypass CDN hotlink protection on Cloudflare-protected sources

Skills

PythonFastAPIPlaywrightBeautifulSoupFletUvicornDockerDocker Composen8n
Manhwa Agent Bot

Project Abstract & Details

An automated offline-first dashboard that tracks personalized reading lists from Asura Scans. Uses a Playwright-based browser emulation layer to detect new unread chapters and displays them in a Flet desktop widget. An n8n workflow orchestrates 6-hourly scans with Docker containerization, automatically opening new chapters in the browser.

Key Features

Offline-First Dashboard

Self-hosted with Docker Compose — fully air-gapped operation with no external service dependencies

Browser Emulation

Playwright-based headless Chromium scraper with User-Agent spoofing and 4-second rate-limit throttling to avoid detection

n8n Automation

12-node n8n workflow with dual-trigger (6-hour cron schedule + on-demand webhook) for tracking reading list updates

Flet Desktop Widget

Cross-platform desktop GUI built with Python Flet — real-time reading list updates with one-click chapter launch

CDN Bypass Pipeline

Base64-encoded image pipeline to circumvent Cloudflare hotlink protection while maintaining visual previews

3-Tier Architecture

Scraper tier (Playwright/FastAPI) → Orchestrator tier (n8n in Docker) → Presentation tier (Flet desktop app)

Architecture

Three-tier containerized architecture: (1) Scraper tier — Playwright headless Chromium controlled via FastAPI REST API scrapes Asura Scans for chapter updates, bypasses Cloudflare with User-Agent rotation and rate-limit throttling. (2) Orchestrator tier — n8n automation platform running in Docker with 12-node workflow triggered every 6 hours or via webhook; compares current chapter list against stored state and triggers notification on new chapters. (3) Presentation tier — Flet desktop app displays reading list with chapter status (read/unread/new) and one-click browser launch.

Architecture diagram placeholder

Outcomes & Metrics

3
Docker Tiers
12
n8n Nodes
6hr
Auto-Scan
0
Cloud Deps

© Developer Portfolio by KV