MoonAI: Introducing Large Language Models to Astronomical Observation—An AI-Driven Tool for Lunar Phase and Crescent Visibility Prediction

MoonAI is an open-source Python tool that leverages large language models (LLMs) like Gemini, Groq, and LLaMA3 to generate lunar phase visibility reports. It aims to provide intelligent solutions for Islamic calendar calculations and astronomical observations, combining modern AI with traditional astronomical needs. Key users include astronomers, Islamic calendar researchers, software developers, and tech enthusiasts interested in astronomical computing.

Lunar phase observation and Islamic calendar calculation have long histories in astronomy and religious practices. Traditionally, determining the visibility of the crescent moon (Hilal) requires complex astronomical calculations and field observation experience. With the development of AI, applying LLM reasoning capabilities to this ancient field has become a new exploration direction, leading to the birth of the MoonAI project.

MoonAI is a Python-based open-source tool designed to generate lunar phase visibility reports. Its core innovation lies in using LLMs' natural language understanding and reasoning to convert complex astronomical data into readable reports. Target users include: astronomers and enthusiasts, Islamic calendar researchers, software developers needing lunar data, and tech personnel interested in astronomical computing.

MoonAI follows a model-agnostic design, supporting multiple mainstream LLMs:

1. Google Gemini: Uses advanced multimodal models for lunar data analysis.
2. Groq: Achieves fast responses via high-performance inference API.
3. LLaMA3 via Ollama: Supports local deployment for data privacy and offline use. Core functional modules: lunar data acquisition (integrating astronomical libraries), visibility analysis (based on location and atmospheric conditions), report generation (LLM converts technical data to structured natural language), and multi-format output.

Islamic Calendar Calculation: Assists in determining important dates like Ramadan and Eid by providing scientific support for crescent visibility. Astronomy Education & Popularization: Natural language reports lower the barrier to understanding astronomical knowledge, suitable for school courses, planetarium activities, and personal learning. Agricultural & Maritime Applications: Structured reports can be integrated into decision support systems for farming and navigation.

Technical Highlights:

• Prompt engineering optimization for astronomical terminology and calculation logic.
• Multi-source data fusion (ephemeris data, real-time weather, geolocation services).
• Scalable architecture (modular design for adding new LLMs, integrating functions, customizing report templates). Open Source Contribution: The project welcomes community contributions, including improving astronomical algorithms, adding more LLM support, optimizing report templates, and improving documentation/examples.

Future Directions:

1. Mobile app development (iOS/Android for on-site observation).
2. RESTful API service for third-party integration.
3. Historical lunar data analysis and visualization.
4. Multilingual support (Arabic, Urdu, etc.). Conclusion: MoonAI demonstrates the innovative application potential of AI in traditional fields. By introducing LLMs to astronomical observation, it simplifies complex calculations and provides new tools for interdisciplinary research, a model worth emulating in other professional fields.