How it works
Enterprise routing is server-side and transparent. The API key prefix determines where data flows:| Key prefix | Infrastructure | Data location |
|---|---|---|
vrin_ | Vrin shared infra | Vrin’s AWS account |
vrin_ent_ | Customer’s AWS account | Customer’s AWS account |
query(), insert(), upload_file(), etc.) are identical for both flows. When you use a vrin_ent_ key, Vrin’s backend:
- Looks up your organization’s infrastructure config in DynamoDB
- Assumes an IAM role in your AWS account via STS
- Routes all operations (Neptune, OpenSearch, S3, Bedrock) to your resources
- Returns results without persisting anything in Vrin’s infrastructure
Data sovereignty guarantees
- Documents uploaded with
vrin_ent_keys are stored in your S3 bucket - Facts are written to your Neptune graph database
- Vector embeddings are indexed in your OpenSearch cluster
- AI model calls use your Bedrock configuration
- Vrin’s infrastructure only handles request routing and authentication
Enterprise components
Your AWS account hosts:| Service | Purpose |
|---|---|
| Amazon Neptune | Knowledge graph (entities, relationships, facts) |
| Amazon OpenSearch | Vector store (embeddings, BM25 index) |
| Amazon S3 | Document storage |
| Amazon Bedrock | LLM and embedding model access |
| Component | Purpose |
|---|---|
| API Gateway + Lambda | Request routing, authentication |
| DynamoDB | Organization config, user accounts |
| STS | Cross-account role assumption |
Getting started
Sign up for enterprise
Contact vedant@vrin.cloud or sign up through the enterprise portal.
Deploy infrastructure
Follow the deployment guide to set up Neptune, OpenSearch, S3, and IAM roles in your AWS account.
Configure Vrin
Save your infrastructure configuration through the enterprise client or web portal.
Next steps
- Enterprise Client — Portal management methods
- Deployment Guide — Step-by-step AWS setup