Skip to content

Roshians/PathShala_AI

 
 

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

81 Commits
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

PathShala AI

AI for Bharat Project Submission Repo
"GitHub Copilot for India's 1.1 Lakh Rural Teachers"

PathShala AI is an AI-powered MVP designed to help rural teachers in India generate structured, multi-grade lesson plans in seconds. A teacher can speak their requirements naturally in Hindi or English (e.g., "Aaj mujhe Class 1 ko vowels sikhane hain aur Class 3 ko multiplication table"), and the system instantly returns a high-quality, practical lesson plan.

Project Context

India has 1,10,000 single-teacher schools where one teacher manages Grades 1–5 simultaneously (60 students, 5 subjects) with no setup, internet, or support staff. Writing 5 lesson plans by hand every evening creates a crushing 2-hour daily administrative burden that drives teacher burnout.

PathShala AI solves this by giving every teacher an AI co-teacher. Unlike generic chatbots, the system encodes multi-grade pedagogic constraints (parallel activities) into an LLM and is accessible directly via Voice or Android App, specifically optimized for rural India's low-resource environments.

🚀 Key Features

  • Multilingual Support: Primary focus on Hindi, with English support.
  • Multi-Channel Delivery:
    • Android App: Voice input, screen display, and Text-to-Speech playback.
    • Phone Call: Direct integration over voice via Twilio — built-in floating dialer FAB with interactive tooltip.
    • WhatsApp: Asynchronous delivery of structured lesson plans directly to the teacher's WhatsApp.
  • Speed & Quality: Generates context-aware, NCERT-aligned, rural-optimized plans in under 15 seconds using powerful reasoning models.

🛠️ Tech Stack

  • Backend Services: AWS Lambda (Python serverless architecture)
  • Frontend: Native Android (Kotlin, Jetpack Compose, Retrofit)
  • AI Engine: Gemini 2.5 Flash / Claude 3.5 Sonnet
  • Communications: Twilio Voice API & WhatsApp API

🧪 Testing the Prototype (For Hackathon Judges)

Since this project currently uses a Twilio Trial account, we cannot send messages or OTPs to unverified numbers. To test the WhatsApp OTP and Lesson Plan delivery features yourself, please join our Twilio WhatsApp Sandbox first:

  1. Open WhatsApp on your phone.
  2. Send the message join become-neighbor to +14155238886 (Twilio's Sandbox Number).
  3. You will receive a confirmation message from Twilio that you have joined the Sandbox.
  4. You can now use the Android App to request an OTP and generate lesson plans using your WhatsApp number!

💻 Setup & Installation

Note: The Python backend has been migrated to AWS Lambda and removed from this repository. This repo now exclusively hosts the Frontend Android App.

To test the application instantly without building:

  1. Go to the Releases page
  2. Download the latest PathShala-AI.apk
  3. Install and run on your Android device.

Frontend Development (Android)

If you wish to build the app from source:

  1. Open Android Studio.
  2. Select Open and choose the frontend/ directory of this repository.
  3. Sync Gradle files.
  4. Connect your Android device or start an emulator.
  5. Click Run to build and install the PathShala AI application on your device.

📖 Additional Context

For more detailed product requirements, user flow, and API contracts, please refer to the MVP PRD: PathShala AI MVP PRD

About

AI for Bharat

Resources

Stars

Watchers

Forks

Packages

 
 
 

Contributors

Languages

  • HTML 84.2%
  • Kotlin 15.7%
  • Java 0.1%