Recipes
All recipes
TypeScriptNext.jsIntermediate

Aria Co-Teacher

An AI-powered co-teacher for virtual classrooms.

Source on GitHubView raw recipe

Recipe prompt

Paste into Cursor, Claude Code, v0, or your coding agent
Use with your coding agent
You are implementing the "Aria Co-Teacher" recipe in this project.

Read the recipe markdown first:
https://raw.githubusercontent.com/saksham456456/Aria/main/README.md

Use the source repository for cross-reference:
https://github.com/saksham456456/Aria

Build this recipe into the user's app using the markdown as the implementation guide. Inspect related source files through the repository links when the recipe points to them. Ask before installing new dependencies.

Recipe

Rendered from the configured recipe markdown.

Raw

Overview about this project: https://youtu.be/0PwnZ5qvT-M #Only judge based on Overview not the demo

Short Demo: https://youtu.be/SxHqx6rH9L8

ARIA: The AI Voice Co-Teacher

Built for the EchoSphere Hackathon

![License: MIT](https://opensource.org/licenses/MIT) ![Version](https://github.com/saksham456456/Aria/releases) ![Next.js](https://nextjs.org/)

ARIA is a real-time, intelligent Voice AI Co-Teacher designed to solve the "silent student" problem in online classrooms. Instead of a teacher lecturing to a wall of muted microphones, ARIA joins the video call as an active participant, helping to explain complex topics, detect learning gaps, and encourage student participation—all in real-time with sub-500ms latency.

✨ Key Features

  • Real-Time Video Classroom: Seamless, low-latency video and audio rooms supporting multiple participants (Teacher & Students).
  • Agora Conversational AI: ARIA is powered natively by the Agora Conversational AI Engine. She possesses deep social awareness, utilizing the Socratic Method to gently guide struggling students without interrupting the main teacher.
  • Live Confusion Radar: An animated, glassmorphic UI radar that actively monitors the classroom for confusion (e.g., "how", "why", "this is hard") and visualizes learning friction in real-time.
  • Auto-Spawning Pop Quizzes: The teacher can instantly deploy AI-generated multiple-choice quizzes based on the last 5 minutes of transcript. Powered by Groq and broadcasted instantly to all students via Supabase Realtime WebSockets.
  • Agent Brain Terminal: A cyberpunk-inspired side panel that visualizes the AI's "internal thoughts", streaming STT intercepts and LLM reasoning logs live to the teacher.
  • Post-Class Analytics: At the end of a session, teachers receive an AI-generated summary of student insights and learning gaps.

💻 Tech Stack

  • Frontend: Next.js 14, React, Tailwind CSS, Lucide Icons
  • Database & Real-Time Sync: Supabase (PostgreSQL + WebSockets)
  • Video/Audio Layer: Agora RTC Web SDK
  • Voice AI Engine (ARIA): Agora Conversational AI SDK (agora-agents) using gpt-4o-mini and MiniMax TTS.
  • LLM Inference: Groq (llama-3.3-70b-versatile) for instant Pop Quizzes and Post-Class Summaries.

🚀 Running Locally

  1. Clone the repository:

``bash git clone https://github.com/saksham456456/Aria.git cd Aria ``

  1. Install dependencies:

``bash npm install ``

  1. Set up Environment Variables:

Create a .env.local file in the root directory and add the following keys: ```env # Supabase (Database & Real-Time) NEXTPUBLICSUPABASEURL=yoursupabaseurl NEXTPUBLICSUPABASEANONKEY=yoursupabaseanonkey SUPABASESERVICEROLEKEY=yoursupabaseservicekey

# Agora (Video & Conversational AI) NEXTPUBLICAGORAAPPID=youragoraappid AGORAAPPCERTIFICATE=youragora_certificate

# Groq (For Pop Quizzes & Summaries) GROQAPIKEY=yourgroqapi_key ``` Note: Ensure that the Conversational AI Engine is toggled ON in your Agora Console so the managed AI models function properly.

  1. Run the development server:

``bash npm run dev ``

  1. Test the Application:

Open http://localhost:3000 in two different browsers (or one Incognito window) to simulate a Teacher and a Student joining the same classroom!

Built with ❤️ for the EchoSphere Hackathon.