Agentic AI Mini-Project

Multi-agent weather assistant with a streaming UI (SSE) and a FastAPI backend.

How to run the project

1) Install prerequisites

Install Node.js + npm

• Use your preferred method (e.g., Node Version Manager, system package manager, or the official installer).
• Verify:
  • node -v
  • npm -v

Install uv (Python package manager)

• Install uv:
  • macOS/Linux: curl -LsSf https://astral.sh/uv/install.sh | sh
  • Windows (PowerShell): irm https://astral.sh/uv/install.ps1 | iex
• Verify:
  • uv --version

2) Backend (FastAPI)

From the project root:

1. Create and sync the Python environment:
   • uv venv
   • uv pip install -r pyproject.toml
2. Add your Azure OpenAI credentials in .env (see api/utils/ai_model.py):
   • AZURE_OPENAI_ENDPOINT
   • AZURE_OPENAI_API_VERSION
   • AZURE_OPENAI_DEPLOYMENT_NAME
   • AZURE_OPENAI_API_KEY
3. Start the API:
   • cd api
   • uvicorn main:app --reload

The streaming endpoint is:

• GET http://localhost:8000/weather?query=<what is the weather in some city?>

3) Frontend (Vite + React)

From the project root:

1. Install dependencies:
   • cd web
   • npm install
2. Start the dev server:
   • npm run dev
3. Build the webpack:
   • npm run build
4. Start the preview server
   • npm run preview

Open the app in your browser (Vite will print the local URL).

Design overview and agent roles

This system uses two cooperating agents:

1. Agent A (Coordinator)

   • Interprets the user’s request.
   • Extracts the location (city) and plans the workflow.
   • Delegates data retrieval to Agent B.
   • Produces the final, user-friendly response.

2. Agent B (Weather Agent)

   • Calls the weather tool with latitude and longitude.
   • Returns the raw weather data to Agent A.

Streaming architecture

• The api streams tokens as Server-Sent Events (SSE) from /weather.
• Vite uses EventSource to render the assistant response incrementally.

Sample inputs and outputs

1. Input: Hi

Output (streamed):

• “Hello! How can I help you today? If you're looking for weather details, please mention the city or location you're interested in.”

2. Input: What is the weather in New Delhi?

Output (streamed):

• “The current weather in New Delhi is as follows:- Temperature: 17.5°C- Conditions: (No specific sky condition available, but it is common for this data to represent clear to partly cloudy skies at this temperature and time.)- Humidity: 58%- Wind Speed: 2.6 km/hLet me know if you need a more detailed forecast or additional weather information!Current weather in New Delhi: The temperature is 17.5°C with 58% humidity and light wind at 2.6 km/h. Conditions are likely clear or partly cloudy.Key metrics:- Temperature: 17.5°C- Humidity: 58%- Wind: 2.6 km/hLet me know if you would like a more detailed report or forecast!”

3. Input: San Francisco

Output (streamed):

• “The current weather in San Francisco is as follows:- Temperature: 9.9°C- Weather Conditions: (No specific condition data)- Humidity: Around 84% (based on recent hourly data)- Wind Speed: 4.1 km/hLet me know if you need more details or a forecast!Current weather in San Francisco:It is currently 9.9°C with high humidity (84%) and a light wind at 4.1 km/h. Weather conditions data was not provided.Key metrics:- Temperature: 9.9°C- Humidity: 84%- Wind: 4.1 km/hIf you'd like more information or a forecast, just let me know!”

Note: Exact values depend on the Open-Meteo API at request time.