AI Web Chat

A local web chat application powered by a Python FastAPI backend and the Gemini CLI. This project allows you to chat with an AI agent in a modern web interface, featuring multi-turn conversation capabilities.

Features

• FastAPI Backend: Robust and fast Python backend serving the application.
• Gemini CLI Integration: Utilizes the gemini CLI for intelligent responses.
• Multi-turn Conversations: Supports maintaining context across multiple exchanges using session management.
• Modern UI: Clean and responsive HTML/CSS/JS frontend.
• Streaming-like Experience: Visual feedback while the AI is processing.

Prerequisites

• Python 3.10+
• Gemini CLI: Ensure the gemini command-line tool is installed and authenticated.
• uv (Optional but recommended): For fast Python package management.

Installation

1. Clone the repository:

   git clone <repository-url>
   cd ai-web-chat

2. Install dependencies:

   Using uv (recommended):

   uv sync

   Or using standard pip:

   pip install -r requirements.txt

   (Note: If requirements.txt is missing, install directly: pip install fastapi uvicorn)

Usage

1. Start the server:

   uv run main.py

   Or with python directly:

   python main.py

   The server will start at http://127.0.0.1:8000.

2. Access the Chat: Open your browser and navigate to: http://127.0.0.1:8000/static/index.html

3. Start Chatting: Type your message in the input box and hit Send. The AI will respond, remembering the context of your conversation.

Project Structure

• main.py: The FastAPI application entry point. Handles chat endpoints and invokes the Gemini CLI.
• static/: Contains the frontend assets (index.html, style.css, script.js).
• pyproject.toml: Project configuration and dependencies.

Troubleshooting

Address already in use

If you see an error like [Errno 98] error while attempting to bind on address ('127.0.0.1', 8000): address already in use, it means another process is already using port 8000.

Solution:

1. Find the process using port 8000:

   lsof -i :8000

   Or using netstat:

   netstat -nlp | grep :8000

2. Kill the process: Take the PID (Process ID) from the output above and run:

   kill <PID>

   (Replace <PID> with the actual number, e.g., kill 12345)

   If the process refuses to close, use -9 to force kill it:

   kill -9 <PID>

3. Retry starting the server:

   uv run main.py