Patmos v2 🚀
An AI-powered research platform An AI-powered research platform leveraging multi-query RAG, vector embeddings (pgvector), and agentic tool-calling to deliver fast, context-aware analysis and semantic search.
🌐 Live Application: patmosresearch.com
📌 Overview
Patmos v2 is engineered for intelligent research automation. By combining semantic search capabilities with vector databases, the platform retrieves relevant document contexts and generates precise, source grounded insights.
It features a production ready SaaS infrastructure with end-to-end user authentication, automated payment subscriptions, and serverless rate limiting.
✨ Key Features
- 🤖 Retrieval-Augmented Generation (RAG): Context-aware question answering powered by semantic vector search and hybrid retrieval.
- ⚡ Agentic RAG & Tool Calling: Orchestrated multi-query vector retrieval and native function-calling using OpenAI (gpt-4o) and Supabase (pgvector) for high-precision, context-bounded responses.
- 🗄️ Vector Storage & Database: Built on Supabase (PostgreSQL) for scalable vector embeddings (
pgvector), relational storage, and Row-Level Security (RLS).
- 🚀 Caching & Rate Limiting: Serverless Redis infrastructure powered by Upstash to handle request throttling and optimize query response times.
- 💳 Monetization & Webhooks: Integrated billing pipeline using LemonSqueezy with automated webhook verification for subscription management.
- 🔐 Authentication: Secure user login and identity management supporting Google OAuth via Supabase Auth.
🛠️ Tech Stack
| Layer | Technologies |
|---|
| Frontend & Framework | Next.js, React, TypeScript, Tailwind CSS |
| AI & Embeddings | OpenAI API |
| Database & Auth | Supabase (PostgreSQL, pgvector, Supabase Auth) |
| Caching & Performance | Upstash Redis |
| Payments & Billing | LemonSqueezy |
| Deployment & Hosting | Vercel |
