An interactive solar system I built where you can click on planets to see their data and ask an AI assistant about ISRO missions. The AI runs locally using LLaMA3 through Ollama — no API keys needed.
Live Demo: https://saketh-salagrama.github.io/Interactive_Solar_System/
Note: The chat assistant needs local setup to work. The planet simulation works directly on the live link.
Languages: Python, JavaScript Frontend: HTML, CSS, JavaScript Backend: FastAPI, LangChain, Uvicorn AI: LLaMA3 (via Ollama), LangChain PromptTemplate Tools: VS Code, Git, Ollama
What it does:
- 8 planets orbiting the sun with different speeds
- Click any planet to see its physical data and related ISRO mission
- Chat with an ISRO-themed AI assistant powered by LLaMA3
- Chat panel can be minimized
How the AI prompt evolved: v1 — Sent raw question directly to Ollama — answers were too generic v2 — Added ISRO scientist persona in the prompt — much more focused v3 — Used LangChain PromptTemplate — clean and structured output
Run Locally:
Install Ollama from https://ollama.com/download, then:
ollama pull llama3
git clone https://github.com/Saketh-Salagrama/Interactive_Solar_System.git
cd Interactive_Solar_System
pip install fastapi uvicorn langchain-ollama langchain-core
# Terminal 1
ollama serve
# Terminal 2
uvicorn app:app --port 5000 --reload
Then open index.html with Live Server in VS Code
Saketh — B.Tech CSE 3rd Year Exploring Generative AI, LangChain and local LLM deployment.