TechScopeAI: Intelligent System for Technical Startup Founders

A production-grade, cloud-native platform that combines RAG (Retrieval-Augmented Generation), multi-agent orchestration, and LLM failover to deliver comprehensive guidance for technical startup founders across pitch creation, competitive analysis, marketing, patents, policies, and team building.

What It Does

TechScopeAI is an AI-powered platform that helps technical startup founders with essential tasks. Instead of generic AI responses, it uses a specialized knowledge base and multiple specialized agents to provide domain-specific, actionable guidance.

Key Components

1. RAG (Retrieval-Augmented Generation) System

• Knowledge Base: Curated datasets from Kaggle, HuggingFace, GitHub, and other sources
• Vector Search: Uses Weaviate Cloud with HNSW indexing for fast similarity search
• Domain Expertise: Separate knowledge bases for competitors, marketing, IP/legal, policies, team building, and pitch examples
• How it works: Agents query the knowledge base before generating responses, ensuring answers are grounded in relevant startup knowledge

2. Multi-Agent Orchestration

• Specialized Agents: Each agent handles a specific domain:
  • Pitch Agent: Creates pitch decks, elevator pitches, and presentation slides. Integrates with Gamma.ai to generate professional presentations with automatic image enhancement via MCP Image Search. Supports multiple themes (startup-pitch, venture-capital, minimalist, modern-tech, executive)
  • Competitive Agent: Analyzes competitors and market positioning
  • Marketing Agent: Generates marketing strategies and content
  • Patent Agent: Conducts patent searches and IP analysis
  • Policy Agent: Drafts privacy policies and legal documents
  • Team Agent: Analyzes team needs and generates job descriptions
• Coordinator Agent: Orchestrates multiple agents for complex tasks
• How it works: Agents use RAG for context, MCP tools for external data, and LLMs for generation

3. MCP (Model Context Protocol) Tools

• External Service Integration: Connects agents to real-time data sources
• Available Tools:
  • Web Search: DuckDuckGo for real-time information
  • Image Search: Pexels/Unsplash for professional images (used by Pitch Agent for Gamma presentations)
  • Patent Search: USPTO database for IP research
  • Content Extraction: Web page content extraction
• How it works: Centralized tool server that agents call when they need external data, avoiding hardcoded API calls

What Founders Get

• Pitch Decks: AI-generated pitch decks with professional visuals via Gamma.ai integration
• Competitive Analysis: Market positioning and competitor insights
• Marketing Strategies: Data-driven marketing recommendations
• IP Guidance: Patent searches and intellectual property advice
• Legal Documents: Privacy policies and compliance documents
• Team Planning: Job descriptions and team structure recommendations

TechScopeAI combines specialized knowledge, real-time data access, and multi-agent orchestration to provide technical startup founders with actionable, domain-specific guidance across all critical business areas.

🌐 Application link

👉 https://techscopeai-web-541402991388.us-central1.run.app/

---

📋 Project Resources

Resource	Link
Live Application	TechScopeAI Web App
Full demo video with explanation	Full Video
Application demo video	Application demo
GitHub Repository	BigDataGroup1/TechScopeAI
Google Codelab	Codelabs
Architecture Diagram	View on Eraser.io

---

🛠️ Technologies

---

🏗️ Architecture Diagram

---

✨ Key Features

Feature	Description
🎯 Pitch Agent	Generate pitch decks, elevator pitches, and investor-ready presentations and gamma ppt
📊 Competitive Agent	Analyze competitors and market positioning
📱 Marketing Agent	Create marketing content and growth strategies
💡 Patent Agent	Assess patentability and IP strategy
📋 Policy Agent	Generate company policies and compliance documents
👥 Team Agent	Team analysis and job description generation
🔄 LLM Failover	Automatic switching between OpenAI GPT-4 and Google Gemini
🗄️ Weaviate RAG	Semantic search across domain-specific knowledge bases
🖼️ MCP Tools	Web search, image search (Pexels), patent search (USPTO)

---

🚀 Installation & Setup

Prerequisites

• Python 3.10+
• React
• Weaviate Cloud
• Google cloud service account

Step 1: Clone Repository

git clone https://github.com/BigDataGroup1/TechScopeAI.git
cd TechScopeAI

Step 2: Environment Variables

cp env.example .env

Edit .env with your API keys:

# LLM (at least one required)
OPENAI_API_KEY=your-openai-key
GEMINI_API_KEY=your-gemini-key

# Weaviate Cloud (required)
USE_WEAVIATE_QUERY_AGENT=true
WEAVIATE_URL=https://your-cluster.weaviate.cloud
WEAVIATE_API_KEY=your-weaviate-key

# Optional
PEXELS_API_KEY=your-pexels-key
GAMMA_API_KEY=your-gamma-key

Step 3: Install Dependencies

# Backend
pip install -r requirements.txt

# Frontend
cd frontend
npm install
cd ..

Step 4: Run Application

Terminal 1 - Backend:

python -m uvicorn backend.main:app --reload --port 8000

Terminal 2 - Frontend:

cd frontend
npm run dev

Step 5: Open Browser

Go to: http://localhost:5173

---

📁 Project Structure

TechScopeAI/
├── backend/              # FastAPI application
│   ├── api/routes/       # API endpoints
│   ├── services/         # Business logic
│   └── main.py           # Entry point
├── frontend/             # React + TypeScript
│   ├── src/pages/        # Page components
│   ├── src/services/     # API client
│   └── src/store/        # State management
├── src/
│   ├── agents/           # AI agents (pitch, marketing, etc.)
│   ├── rag/              # RAG pipeline (Weaviate)
│   └── mcp/              # MCP tools
├── docker-compose.yml
└── requirements.txt

---

👥 Team Contributions

Name	Contribution
Tapas Desai	33.3% - Frontend Development, FastAPI backend, Cloud Run Deployment
Aksh Ashish Talati	33.3% - AI Agents Implementation, Weaviate Cloud , Cloud run deployment , Backend
Swathi Jinka Radhakrishna	33.3% - RAG Pipeline, Weaviate cloud ,MCP integration, Cloud Run Deployment, Backend

---

📝 Attestation

WE ATTEST THAT WE HAVEN'T USED ANY OTHER STUDENTS' WORK IN OUR ASSIGNMENT AND ABIDE BY THE POLICIES LISTED IN THE STUDENT HANDBOOK.

Contribution Declaration:

• All code written collaboratively with clear division of responsibilities
• External libraries and APIs properly attributed
• No code copied from previous course submissions or other teams

---

📄 License

This project is licensed under the MIT License.

---

Made with ❤️ by the TechScopeAI Team