🚀 PINAC Microservices

A robust, scalable microservices demonstration for the PINAC Workspace

📖 Overview

PINAC Microservices is a comprehensive demonstration of a modern, distributed architecture designed to power the intelligent features of the PINAC Workspace. This project showcases how to decouple complex AI and search functionalities into independent, scalable, and secure microservices.

By leveraging Cloudflare Workers for serverless AI inference and Dockerized Nginx for a centralized API Gateway, this project achieves high performance, low latency, and robust security.

🏗️ Architecture

The system is composed of three main components:

1. API Gateway (Nginx): The single entry point for all client requests. It handles routing, load balancing, and enforces security by validating requests against the Auth Service.
2. Auth Service (Node.js): A dedicated service for verifying Firebase JWT tokens, ensuring that only authenticated users can access the AI capabilities.
3. Serverless Workers (Cloudflare):
   • Cloud LLM: Provides direct access to advanced Large Language Models (Google Gemma 2).
   • AI Web Search: Combines Tavily search with AI to provide intelligent, context-aware web search results.

🧩 Services Breakdown

Note

Detailed instructions for each microservice can be found in their individual folder's README file.

1. 🛡️ API Gateway & Auth Service

Located in api_getway/

• API Gateway: An Nginx server configured as a reverse proxy. It routes traffic to the appropriate downstream services based on the URL path (/api/ai, /api/web-search).
• Auth Service: A lightweight Node.js application that validates Authorization: Bearer <token> headers using Firebase Admin SDK (via jose library).

2. 🧠 Cloud LLM

Located in cloud_llm/

• Function: Provides a direct interface to run AI models.
• Features: Handles chat messages and returns AI-generated responses.

3. 🔍 AI Web Search

Located in ai_web_search/

• Function: Performs intelligent web searches.
• Integration: Uses LangChain and Tavily to fetch real-time web data.
• Processing: Summarizes and contextualizes search results using Cloudflare AI.

🔌 Usage

Once deployed, you can access the services via the API Gateway (default port 8080).

1. AI Chat Completion

POST http://localhost:8080/api/ai
Authorization: Bearer <FIREBASE_ID_TOKEN>
Content-Type: application/json

{
  "messages": [
    { "role": "user", "content": "Explain quantum computing." }
  ]
}

2. AI Web Search

POST http://localhost:8080/api/web-search
Authorization: Bearer <FIREBASE_ID_TOKEN>
Content-Type: application/json

{
  "query": "Latest advancements in AI 2025"
}

📄 License

This project is licensed under the MIT License.

---

This project is for demonstration purposes. For production use, review security settings and adapt to your requirements.

Built with ❤️ for the PINAC Workspace