🐾 AI Konkurentų Stebėsenos Sistema (ADK)

AI-powered competitor monitoring and market analysis system built with Google Agent Development Kit (ADK) and Streamlit.

Uses real data — no mocks. ADK agents leverage Google Search grounding and web scraping to collect actual market intelligence.

Architecture

The system uses a multi-agent pipeline built on Google ADK:

1. Competitor Discovery Agent — uses Google Search grounding to find real competitors
2. Price Research Agent — uses Google Search + web scraping to collect actual product prices
3. Data Normalization Agent — AI-powered standardization of product names and prices
4. Deterministic Analysis — Pandas-based median/average calculations, positioning logic
5. Trends Agent — AI market trend analysis with strategic recommendations
6. Data Export — CSV downloads for all collected and analyzed data

Features

• Real Google Search via ADK google_search grounding tool
• Real Web Scraping via custom scrape_website function tool
• Multi-Agent System — 4 specialized ADK agents orchestrated via Streamlit
• No Mock Data — all competitor, product, and pricing data comes from real sources
• Price Positioning — deterministic median-based analysis (Premium / Mid-range / Budget)
• Visual Charts — bar chart comparison of brand prices vs. market median
• CSV Export — download raw data, normalized data, and positioning reports

Prerequisites

• Python 3.10+
• A Google Gemini API key

Setup

1. Clone the repository

git clone <your-repo-url>
cd nd

2. Create a virtual environment

python3 -m venv .venv

3. Activate the virtual environment

Linux / macOS:

source .venv/bin/activate

Windows (PowerShell):

.venv\Scripts\Activate.ps1

Windows (CMD):

.venv\Scripts\activate.bat

4. Install dependencies

pip install -r requirements.txt

Running the App

streamlit run prototipe.py

The app will open in your browser at http://localhost:8501.

Usage

1. Enter your Gemini API key in the sidebar
2. Set the brand name and target market (defaults: Kika / Lietuva)
3. Optionally adjust the product category
4. Click "🚀 Pradėti rinkos analizę" to start the analysis
5. The system will:
   • Discover real competitors via Google Search
   • Research actual product prices from online stores
   • Normalize data for cross-store comparison
   • Calculate statistics, medians, and positioning
   • Analyze market trends and generate strategic insights
   • Provide downloadable CSV reports

Project Structure

nd/
├── prototipe.py        # Main Streamlit + ADK application
├── requirements.txt    # Python dependencies
└── README.md           # This file

Technology Stack

• Google ADK — Agent Development Kit for building AI agents
• Gemini 2.0 Flash — LLM powering all agents
• Google Search Grounding — real-time web search via google_search tool
• BeautifulSoup + lxml — web scraping for detailed price extraction
• Streamlit — interactive web UI
• Pandas — deterministic data analysis