Verse AI
Live in productionBR · São Paulo · Enterprise SaaS
Multi-tenant private ChatGPT for enterprise — 7-layer AWS architecture with RAG over company docs, cryptographic tenant isolation, and a multi-LLM gateway (Anthropic, OpenAI, Bedrock). Built for a Brazilian client, currently live in production.
MULTI-TENANT
Enterprise SaaS
HOSTED ON
AWS (full stack)
STATUS
Live in production
LIVE URL
moovehubia.com.br
Real screens, not mockups
What the system does
Tenant isolation
Each company gets its own private knowledge boundary, cryptographically separated from every other tenant.
Instant answers
RAG over the company's docs delivers grounded answers in seconds — no folder digging, no copy-paste.
Conversational AI
Natural multi-turn chat backed by a multi-LLM gateway (Anthropic, OpenAI, Bedrock, self-hosted).
How it is put together
Client Layer
External APIs · React / Next.js web · Tenant SDK widget
Edge & Networking
CloudFront · Route 53 · WAF · API Gateway
Application Layer
NestJS modular monolith on ECS Fargate w/ ALB
AI / Inference
Bedrock · OpenAI · Anthropic · self-hosted SageMaker
Async Processing
SQS ingestion + embedding queues, Fargate workers, DLQ
Data & Storage
RDS Postgres MultiAZ · Qdrant vector DB · S3 · ElastiCache
Observability & Sec
Cognito · KMS · Secrets Manager · CloudWatch · X-Ray
Stack
- NestJS
- TypeScript
- PostgreSQL
- Qdrant
- Redis
- AWS Cognito
- AWS Bedrock
- OpenAI
- Anthropic
- SageMaker
- ECS Fargate
- S3
- SQS
- CloudFront
- WAF
- Route 53
Where it stands
- Architecture
- Modular monolith, queue-decoupled async
- AI gateway
- Multi-LLM routing + embedding pipeline
- Status
- Actively used. Maintenance & roadmap ongoing.
Need a system like Verse AI?
Tell us what you are trying to ship. The engineer who built this one answers.