Data retention policy
User authentication tokens and workspace data are retained for the duration of the user's active subscription. Conversation context within agent threads is stored temporarily to maintain session continuity and is not retained long-term. User-configured workflows, preferences, and rules are stored persistently until the user deletes them or uninstalls the app. Upon app uninstallation or account deletion, all associated user data is deleted within 30 days.
Data archiving and removal policy
Users can request data deletion by emailing support@hirecaddie.ai. Upon receiving a deletion request, all user data — including stored tokens, workflow configurations, preferences, and conversation history — is permanently deleted within 30 days. When a workspace uninstalls Caddie AI, all workspace-level data is automatically queued for deletion within the same timeframe.
Data storage policy
All application data is stored in Supabase (PostgreSQL) with row-level security policies enforcing per-user and per-workspace data isolation. OAuth tokens are encrypted at rest. No Slack message content is stored persistently — message data is processed in-memory during agent execution and not written to our database. User-configured workflows, rules, and tool connection metadata are stored in our database to enable recurring automations.
App/service has sub-processors
yes
Guidelines for sub-processors
App/service uses large language models (LLM)
yes
LLM model(s) used
Claude Sonnet 4 (Anthropic)
LLM retention settings
Slack message content is passed to Anthropic's API for processing. Anthropic does not retain input/output data for model training per their commercial API terms. No Slack data is used to train or fine-tune any LLM.
LLM data tenancy policy
Single-tenant per request — each API call contains only the requesting user's data. No cross-user or cross-workspace data mixing.
LLM data residency policy
Anthropic API processes data in the United States.