AI Weather (Not in the way you think)

AI-powered weather visualizations – aka when AI meets HTML visualization.

See it in action at https://ai-weather.codetopic.eu! ^^

How did this come to exist?

Yapping section—skip to the next one for the real info.

So, one day, one of my friends sent me a link to this web page. It's a simple web page. All it does is, it shows you nine visualizations of analog clocks made by different AIs. Cool, right? So wacky! ^^

I thought, "Why not make something similar?" And here we are. With a web server that generates these visualizations, but for weather.

What is this?

It's a server that asks various AI models to create HTML/CSS visualizations of current weather data, every hour. The results are displayed in real-time on a web page.

Quick Start

Docker Compose

1. Set your OpenWeather API key

   cp .env.example .env
   # Edit .env with your API key

2. Start ollama service

   docker-compose -f docker/docker-compose.yml up ollama -d

3. Pull Ollama models

   docker exec -it docker-ollama-1 ollama pull llama3.2
   docker exec -it docker-ollama-1 ollama pull qwen3
   docker exec -it docker-ollama-1 ollama pull mistral

4. Start the rest of the services

   docker-compose -f docker/docker-compose.yml up -d

5. Access the application at http://localhost:8000

Docker

Option A: Run Ollama in a Container (Recommended)

# Create a shared network
docker network create ai-weather-network

# Start Ollama container
docker run -d \
  --name ollama \
  --network ai-weather-network \
  -v ollama_data:/root/.ollama \
  ollama/ollama:latest

# Pull models
docker exec ollama ollama pull llama3.2
docker exec ollama ollama pull qwen3
docker exec ollama ollama pull mistral

# Start ai-weather container
docker run -d \
  -p 8000:8000 \
  -e WEATHER__API_KEY=your_key \
  -e OLLAMA__BASE_URL=http://ollama:11434 \
  --network ai-weather-network \
  -v $(pwd)/data:/app/data \
  -v $(pwd)/config/config.yaml:/app/config/config.yaml:ro \
  --name ai-weather \
  docker.io/anty0/ai-weather:latest

Option B: Use Locally Running Ollama

If you already have Ollama running on your host machine:

Linux:

docker run -d \
  -p 8000:8000 \
  -e WEATHER__API_KEY=your_key \
  -e OLLAMA__BASE_URL=http://host.docker.internal:11434 \
  -v $(pwd)/data:/app/data \
  -v $(pwd)/config/config.yaml:/app/config/config.yaml:ro \
  --name ai-weather \
  --add-host=host.docker.internal:host-gateway \
  docker.io/anty0/ai-weather:latest

macOS/Windows:

docker run -d \
  -p 8000:8000 \
  -e WEATHER__API_KEY=your_key \
  -e OLLAMA__BASE_URL=http://host.docker.internal:11434 \
  -v $(pwd)/data:/app/data \
  -v $(pwd)/config/config.yaml:/app/config/config.yaml:ro \
  --name ai-weather \
  docker.io/anty0/ai-weather:latest

Note

On macOS/Windows, host.docker.internal resolves to the host automatically. On Linux, the --add-host flag is required.

For available environment variables, see .env.example. For configuration options, see config/config.yaml.example. The /app/config/config.yaml can be omitted if you configure everything via environment variables.

Development setup

Prerequisites

• Python 3.14+
• Ollama installed and running
• OpenWeather One Call API 3.0 key (get one free here)

Installation

1. Clone the repository

   git clone https://github.com/Anty0/ai-weather.git
   cd ai-weather

2. Set up Python environment

   python3 -m venv venv
   source venv/bin/activate  # On Windows: venv\Scripts\activate
   pip install -r requirements.txt
   pip install -r requirements-dev.txt

3. Configure the application

   cp .env.example .env
   cp config/config.example.yaml config/config.yaml

4. Edit .env with your OpenWeather API key

   WEATHER__API_KEY=your_actual_api_key_here

5. Customize config/config.yaml (optional)

   • Set your location (latitude/longitude)
   • Configure AI models
   • Adjust prompt template

6. Pull Ollama models (if using Ollama)

   ollama pull llama3.2
   ollama pull qwen3
   ollama pull mistral

7. Run the application

   uvicorn aiweather:app --reload

8. Open your browser to http://localhost:8000

Running Tests

pip install -r requirements-dev.txt
pytest --cov=aiweather

Code Formatting

black aiweather/ tests/
ruff check aiweather/ tests/
mypy aiweather/

Adding a New AI Provider

1. Create a new provider class in aiweather/ai/
2. Inherit from AIProvider base class
3. Implement generate_html() and is_available() methods
4. Register provider in AIManager.__init__()
5. Add configuration to config.yaml

Example:

from .base import AIProvider

class OpenAIProvider(AIProvider):
    async def generate_html(self, prompt: str, model_id: str, **kwargs) -> str:
        # Implementation here
        pass

    async def is_available(self) -> bool:
        # Check API availability
        pass

Historical Data

All generated visualizations are saved to the data/ directory:

data/
  2025-11/
    28-14/
      weather.json          # OpenWeather API response
      metadata.json         # Timestamp + models
      Llama_2_7B.html       # Generated HTML
      CodeLlama.html
      Mistral.html
    28-15/
      ...
  2025-12/
    ...

Archives are kept forever by default. You can manually delete old directories if needed.

License

Project is licensed under MIT License—see LICENSE file for details

Credits

• Inspired by AI World Clocks
• Weather Data: OpenWeather One Call API 3.0

Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

Support

If you encounter any issues or have questions, please open an issue on GitHub.