Skip to content

Tech Stack

Detailed breakdown of every technology used in Thamizhi and why it was chosen.

Every technology choice prioritizes:

  1. Open source — full control and transparency
  2. Edge-first — fast, global by default
ComponentTechnologyPurpose
Batch schedulerGitHub Actions (cron)Runs scraping every 30-60 min
RSS scraperPython + Feedparser + BeautifulSoupLightweight feed parsing
Browser scraperPython + PlaywrightJS-rendered page scraping
On-demand scraperCloudflare Browser Run APIReal-time URL scraping
AI extractionGroq API (Llama 3)Structured data from raw HTML
ComponentTechnologyPurpose
API serverCloudflare Workers (TypeScript)All server-side logic
FrameworkHono (TypeScript)Lightweight, edge-optimized
AuthCloudflare Turnstile + JWTBot protection + user sessions
ComponentTechnologyPurpose
LLM inferenceGroq APIArticle processing, fact-checking
ModelsLlama 3, Mixtral (via Groq)Open-source LLMs
Image generationPlaywright (HTML→screenshot)Poster creation
ComponentTechnologyPurpose
Mobile appReact Native + ExpoiOS & Android
Web appReact Native Web (shared code)PWA for desktop
State managementReact Context + ZustandLightweight state
ComponentTechnologyPurpose
API hostingCloudflare WorkersEdge-deployed
Static siteCloudflare PagesDocumentation + web app
CI/CDGitHub ActionsAutomated builds & deploys