NeptuneBlue AI SaaS — RAG · Multi-Tenant · FastAPI
Rania Benotman
What I built
NeptuneBlue is an AI-powered SaaS product I'm building from the ground up as founder of NeptuneBlue Technology Ltd (UK). The product targets domain-specific knowledge management using LLMs grounded in custom data.
Technical architecture
React/TypeScript frontend with a Python FastAPI backend
RAG pipeline using GPT-4o and Qwen3 with pgvector for semantic search over custom knowledge bases
Multi-tenant architecture with organization management, team access controls, and role-based permissions
Stripe billing with subscription tiers
Full Docker containerization, Grafana production monitoring, and GitHub Actions CI/CD pipelines
What makes it complex
Building multi-tenant SaaS solo means owning every layer — auth, billing, AI pipeline, infra, and frontend. The RAG implementation required careful chunking strategy, embedding model selection, and retrieval tuning to get context-aware responses that actually ground correctly in the knowledge base rather than hallucinating.
Stack
React · TypeScript · Python · FastAPI · pgvector · Docker · GitHub Actions · Stripe · Grafana
Like this project
Posted May 1, 2026
Multi-tenant AI SaaS with RAG pipeline (GPT-4o + Qwen3 + pgvector), React/TypeScript frontend, Python FastAPI backend, Stripe billing, Docker, Grafana & CI/CD.
Likes
2
Views
6
Clients
NeptuneBlue Technology Ltd
