Gemini Autonomous Agent Demo

This project is a Flask-based web application demonstrating an autonomous agent powered by Google's Gemini models. The agent is capable of performing various tasks through tool integration, including real-time web search, email management, and news aggregation.

Features

• Interactive Chat Interface: Seamless conversation with the Gemini agent.
• Autonomous Capabilities:
  • Web Search: Retrieves real-time information using Google Custom Search.
  • Email Management: Reads unread emails and sends new emails via Gmail integration.
  • News Aggregator: Fetches and summarizes the latest news from Google News RSS feeds.
• Chat Management: Create, load, and manage chat sessions with persistent history.

Tools & Technologies

• Backend: Flask (Python)
• AI Model: Google Gemini (via google-genai SDK)
• Search: Google Custom Search JSON API
• Email: SimpleGmail
• News Extraction: feedparser, newspaper3k, googlenewsdecoder
• Frontend: HTML/CSS (Jinja2 templates)

Setup & Installation

Prerequisites

• Python 3.8+
• A Google Cloud Project with the following APIs enabled:
  • Gemini API
  • Custom Search API
  • Gmail API

Installation

1. Clone the repository:

   git clone <repository-url>
   cd Gemini-Autonomous-Agent-Demo

2. Create and activate a virtual environment:

   python3 -m venv .venv
   source .venv/bin/activate  # On Windows use `.venv\Scripts\activate`

3. Install dependencies:

   pip install -r requirements.txt

Configuration

1. Environment Variables: Create a .env file in the root directory (use .env.example as a reference) and add your API keys:

   GEMINI_API_KEY=your_gemini_api_key
   GEMINI_MAIN_MODEL=gemini-2.0-flash-exp
   GEMINI_QUICK_MODEL=gemini-2.0-flash-exp
   SEARCH_API_KEY=your_google_custom_search_api_key
   SEARCH_ENGINE_ID=your_search_engine_id
   PRIMARY_USER_MAIL=your_email@gmail.com

2. Gmail Credentials:

   • Download your client_secret.json from the Google Cloud Console (OAuth 2.0 Client IDs).
   • Place it in the gmail_credentials/ directory.
   • Run the initial setup script (if applicable) or ensure the app handles the initial OAuth flow to generate gmail_token.json.

Usage

1. Run the application:

   python app.py

2. Access the interface: Open your browser and navigate to http://127.0.0.1:7777.

Project Structure

• app.py: Main entry point for the Flask application.
• routes.py: API route definitions.
• views.py: Logic for chat handling and interactions.
• utils.py: Implementation of tool functions (Search, Email, News).
• tool_declarations.py: Definitions of tools exposed to the Gemini model.
• templates/: HTML templates for the frontend.