Data retention policy
We retain data only as long as needed for business/contractual/regulatory purposes, then securely delete or archive it; PII is deleted or de-identified when it no longer has a business use. Customer data is deleted within 60 days of contract termination.
Data archiving and removal policy
We follow a documented retention schedule and delete or de-identify PII when it’s no longer needed, with customer data deleted within 60 days of contract termination (and deletion cascaded to backups/archives once retention obligations are met).
Data storage policy
Data is classified, access-controlled, and encrypted in transit (TLS) and at rest (AES-256), with encrypted, segregated backups and annual restore tests. We retain data only as needed and delete customer data within 60 days of contract termination; production data is not used in non-prod.
Data center location(s)
United States
Data hosting details
Cloud hosted
Data hosting company
AWS, Azure, GCP
App/service has sub-processors
yes
Guidelines for sub-processors
App/service uses large language models (LLM)
yes
LLM model(s) used
Google Cloud (Gemini, Anthropic Claude), AWS Bedrock (Anthropic Claude)
LLM retention settings
We do not train models and do not store customer prompts or outputs for training with Google Gemini or Anthropic Claude.
LLM data tenancy policy
By default, Levenza processes LLM requests in US regions across Google Cloud and AWS Bedrock, and does not allow training of provider models on our customers' data.
LLM data residency policy
Levenza accesses Google Gemini via Google Cloud (Vertex AI), and Anthropic Claude via Google Cloud (Vertex AI) and AWS Bedrock. We do not train on customer data and do not commingle customer data for model training.