Privacy Policy
Last Updated: June 3, 2026
1. Information We Collect
We collect system metrics, API request contexts, and cognitive grounding facts uploaded to train your assistants (such as text guidelines or internal reference documentation). User conversations processed by the assistant runtime are temporarily held to maintain chat context.
2. How We Use Information
Information is used solely to generate assistant responses, track usage counts in our billing system, and monitor NATS message flows. Grounding documentation files are processed into secure database vector databases to support RAG query searches.
3. Data Protection and Vaults
All customer integration keys and credentials are encrypted using master key wrappers (`platform.secrets` table envelope encryption). Grounding document access limits are protected to prevent cross-tenant data leaks.
4. Sharing Restrictions
We do not sell user data. Conversation summaries or ingestion files are only processed through LLM API protocols to perform responses, with zero data retention policies on model providers where available.