PromptTheStars 🌌

PromptTheStars is an open-source, production-grade AI astrology web application. Powered by Vedic Astrology (Jyotish) and cutting-edge Retrieval-Augmented Generation (RAG), it provides hyper-personalized astrological readings, Macro Birth Time Rectification (BTR), and long-term agentic memory via a daily journaling system.

🚀 Features

• Master User Dashboard: Configure your foundational planetary profile once (DOB, Time, Location).
• Contextual Memory (RAG): The AI remembers your daily journal entries, mapping your human emotions and events directly to astrological transits.
• Dynamic Checklists & BTR: The system schedules future check-ins based on its predictions. Over time, it uses this mathematical feedback loop to subtly refine your birth time accuracy.
• Seamless Local Tool Calling: Uses standard Model Context Protocol (MCP) to compute 100% accurate ephemeris charts on the fly.

🛠 Tech Stack

Frontend (Client)

• Angular 21 (Served via Node 24)
• Angular Material + Customizable Chat Component
• Vanilla CSS with modern aesthetics (glassmorphism, smooth micro-animations).

Backend (Orchestration)

• Java 25 & Spring Boot 4.0.5
• Spring AI (Orchestrates LLM generation and RAG retrieval)
• REST APIs & Model Context Protocol (MCP)

Database & Intelligence

• PostgreSQL + pgvector (Relational data and Vector Embeddings)
• rishi-ai-mcp (Python server providing specialized chart and Dasha generation)
• LLMs: OpenAI / Gemini / Claude (configurable natively via Spring AI)

⚙️ Setup Instructions

Because the application leverages vector databases and Python AI processes communicating with the JVM via native I/O streams, the entire architecture is fully containerized for seamless "one-click" local deployment.

1. Prerequisites

Ensure you have the following installed on your machine:

• Docker
• Docker Compose

Note: You do NOT need to manually install Java, Node.js, Python, Maven, PostgreSQL, or rishi-ai-mcp. Everything is securely built and isolated during the multi-stage Docker builds.

2. Local Deployment

From the root directory of the project, launch the stack:

docker-compose up --build

Docker will automatically:

1. Spin up the PostgreSQL database (with the pgvector extension pre-installed).
2. Build and launch the Spring Boot API (packaging Python internally and orchestrating the rishi-ai-mcp library).
3. Build the Angular Frontend cleanly and serve it over an Nginx web server proxy.

The application and dashboard will be immediately accessible at http://localhost:4200.