Infrastructure & Scalability
Enterprise-grade infrastructure with GCP hosting, multi-tenant architecture, and scalable deployments.
Infrastructure Overview
Craveva AI is built on enterprise-grade infrastructure hosted on Google Cloud Platform (GCP). Our multi-tenant architecture ensures complete data isolation while providing unlimited scalability. The platform consists of two main services: Next.js frontend (port 3000) and Express.js backend API (port 5000), all orchestrated via Docker Compose with Nginx as reverse proxy.
Google Cloud Platform Hosting
Rendering via Kroki...
Usually takes 1-2 seconds
Cloud Storage
Google Cloud Storage for scalable file uploads and data storage
Compute Engine
High-performance virtual machines for backend and frontend services
VPC Network
Secure, isolated network infrastructure with private networking
Multi-Tenant Architecture
Rendering via Kroki...
Usually takes 1-2 seconds
Company-Level Isolation
Each company's data is completely isolated at the database level using company_id filtering on all queries
Location-Level Deployment
Agents can be deployed company-wide or to specific outlets with outlet-level configuration and data filtering
Data Separation
Complete data separation ensures no cross-tenant data access. All MongoDB collections include company_id indexes
Independent Billing
Each company has independent credit balance, usage tracking, and billing rules with Stripe integration
Scalability Features
10+ Outlets
Handles enterprises with 10+ outlets per company with outlet-level isolation
Unlimited Users
Support unlimited users per company with 6 role-based access levels
Unlimited Agents
Create and deploy unlimited AI agents (12 agent types available)
Unlimited Data Sources
Connect unlimited data sources per company (12 DB types + 8 POS + 5 Delivery + Google Drive + Files)
Real-time Processing
Live data synchronization and responsive agent interactions
High Availability
Redundant systems and operational monitoring for reliability
Performance Metrics
Responsive agent interactions powered by optimized query execution
Reliability-focused architecture with redundancy
Designed to support multi-team organizations and deployments
Upcoming Performance Improvements
Beta V2.0 (Feb 14, 2026)Beta 2.0 (Feb 14 - Mar 13, 2026) will introduce significant performance improvements to handle 1000+ concurrent requests with horizontal scaling.
RAG & Caching
- 10-100x performance improvement
- 70-80% faster RAG queries
- Multi-layer caching (CDN, Redis, In-Memory)
Queue System & Scaling
- BullMQ queue system
- Horizontal scaling with PM2 cluster mode
- Auto-scaling based on CPU/memory metrics
- Support for 1000+ concurrent requests
Enterprise Security
Built with security and privacy as core principles. Multi-tenant isolation, encryption, and role-based access control.