🤖 AskVES — Unofficial AI Assistant for VESIT Students

⚠️ Unofficial student project. Not affiliated with or endorsed by VESIT.

Live Demo: ( currently getting a UI overhaul )

AskVES is a RAG-based campus chatbot that answers real-time questions about VESIT college using structured campus data and LLM reasoning.

What It Can Do

• Canteen — prices, menu, budget recommendations
• Teachers — real-time location based on current timetable slot
• Events — upcoming college events and workshops
• Xerox — compare AC vs Non-AC shop speed and variety
• Vending Machine — what's available near hostel 24/7
• Community — Reddit-style discussion board with @ves.ac.in auth, upvoting, and auto-moderation.

Architecture & Tech Stack

• LLM: Groq (Llama 3.3 70B) → Groq (8B fallback) → Gemini 2.0 Flash
• Backend: Modular Python HTTP Server (http.server)
• Database: Supabase (PostgreSQL) + Local JSON fallback
• Frontend: Vanilla HTML/JS/CSS (Chat interface, Admin Dashboard, Community Board)
• Deployment: Railway.app

Project Structure

app/
  main.py            # Main server setup
  routes/
    webhook.py       # WhatsApp Twilio webhook logic
  services/
    ai.py            # MultiAIProvider (Groq/Gemini) fallback logic
    db.py            # Supabase / JSON data loaders
    voice.py         # Voice processing (Upcoming)
  core/
    intents.py       # Query expansion and RAG context building
    router.py        # HTTP route handlers & endpoints
    state.py         # Global state & analytics
  utils/
    logger.py        # Logging helpers
    helpers.py       # General utilities (e.g., slot detection)

Run Locally

git clone https://github.com/000wahab000/AskVES
cd AskVES
pip install -r requirements.txt
# Add .env with GROQ_API_KEY, GEMINI_API_KEY, and SUPABASE_URL/KEY
py main.py
# Open http://localhost:8000

Roadmap & Upcoming Features

✅ Completed Milestones

• Admin Dashboard: Live editing of campus data and system metrics.
• WhatsApp Integration: Twilio webhook routing for mobile access.
• Reddit-Style Ecosystem: Community board with upvotes, automated moderation, and fact-reporting.
• Modular Architecture: Clean app/ folder structure for scalable development.
• increased the response speed: the speed has the been increased to currently at 400ms to 200ms

Massive Data Expansion

• Demographics: Total student counts across B.Tech, M.Tech, Management, Law, and other departments.
• Academics: Detailed breakdown of all B.Tech fields and specializations.
• Placements: Comprehensive placement statistics, salary packages, and top recruiting companies.
• Facilities: Detailed information on all campus facilities.
• Administration & Faculty: Teacher salaries, internal management structure, legal/business info, and "How to get a job here" guides.

Advanced Bot Capabilities

• Voice Integration: Add voice inputs and outputs to make the bot more accessible.
• Multilingual Support: Support queries and responses in multiple regional languages.
• Improved Precision: Fix vague responses by tightening RAG context and system prompts.

Community & UI Improvements

• Dual Authentication System: Differentiate between verified @ves.ac.in students (full access) and standard Gmail users (guest/read-only access).
• Notice Board UI: Evolve the design into 2 main parts, incorporating a Live Notice Board alongside the chat interface.
• Notes Marketplace: A dedicated section for seniors to share or sell study notes.
• Visual Excellence Redesign: Overhaul the interface with modern glassmorphism, dynamic micro-animations, and a curated premium dark-mode color palette.

Built At

UniMerge 1.0 Hackathon — April 2026