Integrating Multi-Tenant Application with Vanna.AI for User Queries

[sinjup]

We have an application with a multi-tenant architecture, and we would like to integrate it with Vanna.AI to allow users from different tenants to ask questions and get responses.
However, we need some guidance on the best practices for setting up this integration in Vanna.AI, and whether Vanna.AI can support multi-tenant environments.

Specifically, we would like to know:

Tenant Isolation: How can we ensure that users' data is isolated between tenants? Are there specific configurations or patterns in Vanna.AI that support multi-tenancy?

User Queries: How can we handle user queries in such a way that each tenant's users can ask questions and receive relevant answers without data leakage?

Training Data: What is the recommended approach for providing tenant-specific data for training purposes? Should we upload data separately for each tenant or use a global dataset with tenant-specific tagging?

Function Calls: Since our application involves tenant-specific logic, what is the best way to integrate tenant-specific functions or stored procedures into Vanna.AI’s training and query mechanisms?

Authentication & Authorization: How can we securely authenticate and authorize users from different tenants to access the Vanna.AI system while ensuring data security and privacy?

Support for Multi-Tenant Setups: Does Vanna.AI support multi-tenant environments where each tenant's data and queries are isolated and handled separately? If not, what would be the best approach to achieve this integration?

@zainhoda