explaingit

bb1nfosec/vaathi

★ 16PythonAudience · generalComplexity · 3/5ActiveLicenseSetup · moderate

TLDR

A free cybersecurity learning site for Indian students in six languages, where each user brings their own LLM API key to keep hosting costs at zero.

Mindmap

mindmap
  root((vaathi))
    Inputs
      User API key
      Student answers
      Topic choice
    Outputs
      Personalized roadmap
      Quizzes and CTFs
      Spaced repetition reviews
    Use Cases
      Learn security in Tamil or Hindi
      Practice CTF challenges
      Track progress with SM-2 review
    Tech Stack
      Next.js
      TypeScript
      Prisma
      Turso
      Vercel

Things people build with this

USE CASE 1

Learn cybersecurity in Tamil, Hindi, or four other Indian languages

USE CASE 2

Run a free hosted CTF arena keyed to your own LLM provider

USE CASE 3

Track topic mastery with SM-2 spaced repetition

USE CASE 4

Deploy a personalized study app for a class without paying for inference

Tech stack

Next.jsTypeScriptPrismaTursoVercel

Getting it running

Difficulty · moderate Time to first run · 30min

Needs a free Groq, OpenRouter, or Together AI key plus Vercel and Turso accounts before the site can serve any chat.

MIT, do what you want, keep the copyright notice.

In plain English

Vaathi is a free cybersecurity learning website aimed at students in India. The pitch in the README is that paid platforms like TryHackMe or HackTheBox cost around fourteen dollars a month and are only in English, while Vaathi costs nothing and runs in six languages: Tamil, Hindi, Telugu, Malayalam, Kannada, and English. The repo has 14 stars and is listed as a Python project, although the actual web app is built with Next.js and TypeScript. The trick that keeps it free is what the author calls Bring Your Own LLM. The student creates an account on a service such as Groq, OpenRouter, or Together AI, picks up a free API key, and pastes it into Vaathi. The hosted Vaathi web app then talks to that key on the student's behalf. Hosting runs on the free tier of Vercel, the database is the free tier of Turso, and the LLM calls are billed to the student's own provider key, which in Groq's case is free with no card. The learning flow has four phases. First the AI asks open ended cybersecurity questions and the student answers in their own words, which the model uses to score depth and to draft a personalized roadmap. Second comes guided learning with explanations, multiple choice quizzes, and small five minute tasks like spotting a bug in a code snippet, decoding Base64 or hex, reading a log excerpt for an attack, or explaining what an nmap command does. Third is a CTF arena where the AI generates challenges that scale with the student's tier. Fourth is review: completed topics are scheduled through the SM-2 spaced repetition algorithm so they come back at growing intervals, with harder ones returning sooner. Around all that sit a Guru AI free form chat that adapts to the student's level, a daily streak counter, a tier system that goes from Egg up to Burn, and badges. The README walks through both a one command deploy.sh script that signs the user into Turso and Vercel and a manual path with prisma db push and Vercel imports. The license is MIT.

Copy-paste prompts

Prompt 1
Run the deploy.sh script to push Vaathi to Vercel and Turso under my own account, and tell me what env vars I still need to set
Prompt 2
Swap the LLM provider from Groq to OpenRouter and update the Bring Your Own LLM settings page
Prompt 3
Add a seventh language to the existing six by translating the UI strings and quiz prompts
Prompt 4
Replace the default tier names from Egg to Burn with a custom tier scheme for my school
Prompt 5
Wire Prisma to a local Postgres instead of Turso and run prisma db push for development
Open on GitHub → Explain another repo

Generated 2026-05-22 · Model: sonnet-4-6 · Verify against the repo before relying on details.