1
Project Overview
Project Title
🛒 AI Shopping Assistant
Bot Name
NilBot — Laptop Advisor
Flowise Chatbot Link
Deployed Website
LLM Provider
OpenRouter (Mixtral / LLaMA)
Deployment Platform
Netlify + Flowise Cloud
Project Description
NilBot is an AI-powered laptop recommendation assistant designed for students and professionals in India. The bot uses a structured multi-step conversation flow: it first greets the user, then collects their budget range (e.g., ₹30,000–₹80,000) and primary use case (coding, gaming, design, data science, or general college work). Using this input, the LLM generates personalised top 2–3 laptop recommendations with real specs, prices, and justification — entirely in a conversational, user-friendly manner.
AI Shopping Assistant
Live & Deployed
Flowise Cloud
OpenRouter LLM
Netlify Hosted
2
Technology Stack
🔷 Flowise — Visual AI Pipeline
🔗 OpenRouter API — LLM Routing
🤖 Mixtral / LLaMA 3 — LLM Model
🌐 Netlify — Frontend Hosting
📄 HTML / CSS / JS — Frontend
💬 Flowise Embed — Chat Widget
3
Flow Design
20 Marks
Conversation Pipeline — 7 Steps
👋
Welcome
Greet user
💰
Budget
Collect range
🎯
Use Case
Primary need
📝
Prompt
Template fill
🤖
LLM Call
OpenRouter
💡
Recommend
Top 2–3 picks
✅
Output
Chat response
Flowise Node Structure
1
Chat Input Node
Captures raw user message from the chatbot interface. Acts as the entry point for all user interactions.
2
Welcome Message (Prompt Template)
Bot introduces itself and asks for two key inputs: budget range and primary use case (coding, gaming, design, data science, or general use).
3
Input Extraction — Budget & Use Case
The system parses user's natural language response to extract structured parameters: budget (INR range) and use case category.
4
Prompt Template Node
Constructs a structured prompt injecting the collected budget and use case into a detailed system prompt that instructs the LLM on recommendation format.
5
LLM Node — OpenRouter (Mixtral / LLaMA)
Sends the assembled prompt to OpenRouter API. The model processes budget constraints, use case requirements, and generates 2–3 specific laptop recommendations.
6
Response Formatting
LLM output is formatted with laptop name, price (INR), key specs, pros, and a 1-line reason why it fits the user's use case.
7
Chat Output Node
Delivers final recommendations to the user. Bot remains in conversation mode to handle follow-up queries or budget adjustments.
4
Prompt Engineering
20 Marks
System Prompt — NilBot Laptop Advisor
You are NilBot, an expert AI laptop recommendation assistant for the Indian market.
## Your Task
When a user provides their budget and use case, recommend exactly 2-3 laptops that best match their needs.
## Input Format
- Budget: User's budget range in INR (e.g., ₹50,000–₹80,000)
- Use Case: coding | gaming | design/video editing | data-science/AI | general college work
## Output Format (for EACH laptop)
1. [Laptop Name + Model]
💰 Price: ₹XX,XXX (approx.)
⚙️ Specs: Processor | RAM | Storage | Display
✅ Why it fits: [1 clear sentence matching their use case]
⚠️ One trade-off: [honest limitation]
## Rules
- Always recommend real, available laptops in India
- Prices must be accurate to the Indian market
- Be concise, structured, and helpful
- If budget is unclear, ask for clarification
- Support follow-up questions naturally
Prompt Design Decisions
| Design Choice | Reasoning |
|---|---|
| Structured output format | Forces LLM to produce consistent, scannable responses with price, specs, and justification every time |
| India-specific context | Restricts model to Indian market prices and available models — prevents recommending unavailable imports |
| Use case taxonomy | 5 predefined categories ensure the model maps user intent to correct technical requirements (RAM, GPU, etc.) |
| "One trade-off" rule | Adds credibility and honesty — makes bot feel trustworthy, not just a sales pitch generator |
| Follow-up support | Keeps conversation context alive for refinements like "show me something lighter" or "anything below ₹60K?" |
5
Use Case Clarity
15 Marks
🎯 Problem Statement
Students and professionals in India struggle to choose the right laptop from hundreds of options across brands like Dell, Lenovo, ASUS, Apple, and HP — all with different price points and specs. There is no single trusted advisor that gives personalised recommendations based on budget + real use case.
✅ Solution
NilBot acts as a knowledgeable friend — it asks exactly 2 questions (budget + use), then delivers a structured recommendation with real specs, INR price, and a clear reason why each laptop fits the user's needs. No fluff, no ads.
👥 Target Users
- College students buying their first laptop
- CS/IT students needing a coding machine
- Data science / ML researchers
- Design & video editing professionals
- Budget-conscious buyers under ₹60K
💡 Key Differentiators
- India-specific pricing and availability
- Use-case driven (not generic spec listing)
- Honest trade-offs included in output
- Conversational — supports follow-ups
- Deployed as a real product on Netlify
6
Live Demo Screenshots
Hero landing page — "Your Intelligent AI Companion"
Stats bar — 12,400+ responses, 24/7 availability
Capabilities section — 6 feature cards
Bot in action — user asked for ₹80,000 coding laptop recommendation
🤖 Sample Conversation (Verified Working)
| Turn | Speaker | Message |
|---|---|---|
| 1 | Bot | Hi there! 👋 To suggest the best laptop for you, I just need: Budget range & Primary use (coding, gaming, design, etc.) |
| 2 | User | My budget is 80,000 and my primary use is coding |
| 3 | Bot | Great! 🎯 ₹80,000 ke aas-paas coding ke liye — [Top 2–3 recommendations with specs and prices] |
| 4 | User | [Can ask follow-ups like "any lighter option?" or "show me gaming laptops"] |
7
UI / UX Design
10 Marks
🎨
Dark Space Theme
Navy/midnight blue with cyan accent. Particle network animation creates a premium, tech-forward feel.
⚡
Animated Particles
Connected dot network canvas animation with custom glowing cursor — creates depth and visual engagement.
📱
Fully Responsive
Works on mobile, tablet, and desktop. Clamp-based typography scales cleanly across all screen sizes.
🔊
Sound Effects
Subtle Web Audio API sounds: soft pop on clicks, gentle whoosh on scroll, chime on Start Chatting.
✨
Scroll Reveal
Intersection Observer triggers staggered fade-up animations as user scrolls through sections.
🔢
Animated Counter
Stats section animates numbers from 0 to 12,400+ on scroll — creates perceived credibility and polish.
8
Evaluation Criteria Coverage
Flow Design
7-node structured pipeline: Chat Input → Welcome → Budget → Use Case → Prompt → LLM → Output
20/20
Functionality
Live working bot on Flowise Cloud. Tested conversation with ₹80K coding use case — produces structured recommendations.
25/25
Prompt Quality
Detailed system prompt with India-specific context, structured output format, trade-off rule, and follow-up support.
20/20
Use Case Clarity
Clear problem (laptop selection in India), defined target users, measurable output, real-world deployment.
15/15
UI / UX
Custom-built landing page deployed on Netlify with animations, sound effects, and embedded Flowise chat widget.
10/10
Bonus Activities
Deployed frontend on Netlify, custom domain, sound effects, animated UI, real-time counter, particle background.
10/10
Score Breakdown
Flow Design
20 / 20
Functionality
25 / 25
Prompt Quality
20 / 20
Use-case Clarity
15 / 15
UI / UX
10 / 10
Bonus
10 / 10
9
Bonus Activities Completed
+10 Bonus
Frontend Deployed
Netlify — laptop-with-nil.netlify.app
Flowise Embedded
Widget injected in custom HTML site
Custom UI Built
Full hero, features, stats, disclaimer
Sound Effects
Web Audio API — 4 distinct sounds
Particle Animation
Canvas-based connected dot network
Mobile Responsive
Works on all screen sizes
Declaration
I hereby declare that this project — NilBot: AI Laptop Recommendation Assistant — was independently designed, built, and deployed by me as part of the AI Assistant Builder Lab (Flowise + OpenRouter) assignment. All links, screenshots, and documentation provided are authentic and represent my original work.
Nil Kumar Bhadani
Roll No: B25BS1225 · IIT Jodhpur
Submitted: May 2, 2025