Business Idea Validator — AI Agent

An AI-powered startup validation platform built with Google ADK and Gemini, deployed on GCP Cloud Run. Analyzes business ideas across market potential, risk, competition, financial viability, and geographical fit for the Indian market — returning structured reports through a conversational chat interface.

Live Demo: https://business-idea-validator-du6h2ejera-el.a.run.app

---

Tech Stack

Component	Technology
Agent Framework	Google ADK
LLM	Gemini 2.5 Flash
Backend	FastAPI + Uvicorn
Memory	In-Memory Session Manager
Deployment	GCP Cloud Run
Containerization	Docker
Data Validation	Pydantic v2

---

Architecture

Browser (Chat UI)
        |
        | HTTP REST
        v
FastAPI Backend (Cloud Run)
        |
        +---> Google ADK Agent
        |           |
        |           v
        |     Gemini 2.5 Flash
        |
        +---> In-Memory Session Manager
                    |
                    v
              Conversation History
              (TTL per session)

The agent operates in two modes determined by session state:

• Analysis Mode — triggered on first message, returns a full structured JSON validation report
• Follow-up Mode — triggered on subsequent messages, returns a conversational consultant response grounded in the original analysis

---

Features

Core Agent Capabilities

• Viability scoring (0–10 float) with honest, calibrated reasoning
• Market analysis across Indian Tier 1, Tier 2, and Tier 3 cities
• Risk, opportunity, and competitive landscape assessment
• Financial viability and scalability evaluation
• Actionable first step recommendations
• Geographical context: UPI ecosystem, Jio penetration, MSME schemes, Startup India

Technical Highlights

• Dual-mode prompt architecture — structured JSON analysis and natural conversation from a single agent
• Session state flag (has_analysis) for reliable mode switching without extra API calls
• Auto-labeled conversation history in the sidebar
• Graceful ADK fallback to direct Gemini if ADK initialization fails
• 6 REST endpoints with Swagger auto-documentation at /docs

---

Project Structure

Local-Business-Idea-Validator/
├── main.py              # FastAPI app, endpoints, middleware
├── agent.py             # VentureCheck ADK agent, prompt engineering, dual-mode logic
├── adk_agent.py         # Standalone ADK agent definition and demo
├── memory.py            # In-memory conversation manager with session state
├── models.py            # Pydantic request/response schemas
├── redis_client.py      # Redis conversation manager (optional)
├── Dockerfile           # Container configuration
├── requirements.txt     # Dependencies
├── .env.example         # Environment variable template
└── UI_Pages/
    ├── index.html       # Chat interface structure
    ├── index.css        # Styling
    └── index.js         # Client logic, auto-conversation creation, sidebar

---

API Endpoints

Method	Endpoint	Description
GET	/health	Service and memory status
POST	/conversation/start	Create a new conversation session
POST	/chat/message	Send message, get validation or follow-up
GET	/chat/history/{id}	Retrieve full conversation history
DELETE	/conversation/{id}	Delete a conversation
POST	/validate	Single validation without session memory
GET	/docs	Swagger interactive API documentation

URL of the Project Deployed on the cloud run

• Note Same features are going to be implemnt in the project in few times
• Url:- https://business-idea-validator-626133859913.asia-south1.run.app

Sample Request

curl -X POST https://business-idea-validator-du6h2ejera-el.a.run.app/chat/message \
  -H "Content-Type: application/json" \
  -d '{
    "idea": "A hyperlocal organic vegetable delivery app for Tier 2 cities",
    "conversation_id": "your-conversation-id",
    "input_type": "new_idea"
  }'

Sample Response

{
  "title": "Hyperlocal Organic Veg Delivery — Tier 2",
  "score": 7.2,
  "verdict": "Promising",
  "market": "Health-conscious urban households in Tier 2 cities...",
  "risk": "Cold chain logistics and consistent supply from local farmers...",
  "opportunities": "Growing health awareness post-COVID and rising disposable income...",
  "competition": "BigBasket and Blinkit dominate Tier 1 but have limited Tier 2 presence...",
  "first_step": "Partner with 2-3 local organic farms this week and run a WhatsApp pilot...",
  "summary": "This idea has real traction potential in underserved Tier 2 markets..."
}

---

Screenshots

---

Local Setup

Prerequisites: Python 3.11+, Google Gemini API key

# Clone and enter directory
cd Local-Business-Idea-Validator

# Create virtual environment
python -m venv env
env\Scripts\activate        # Windows
source env/bin/activate     # macOS/Linux

# Install dependencies
pip install -r requirements.txt

# Configure environment
cp .env.example .env
# Edit .env and add: API_KEY=your_gemini_api_key_here

# Run
python main.py

Open http://127.0.0.1:8000 in your browser.

---

Cloud Run Deployment

gcloud run deploy business-idea-validator --source . --platform managed --region asia-south1 --allow-unauthenticated --set-env-vars API_KEY=your_key_here --port 8000

---

Environment Variables

Variable	Required	Description
API_KEY	Yes	Google Gemini API key from AI Studio
REDIS_URL	No	Redis connection URL for persistent memory
LOG_LEVEL	No	Logging level (default: INFO)

---