Data retention policy
Hubi, operated by Tidio retains customer account data and AI conversation history (messages exchanged directly
with the Hubi assistant) for the duration of the customer's active subscription plus 5 years, after which it is permanently
deleted. Slack OAuth tokens are stored encrypted and retained as long as the workspace integration remains active. Hubi does not
store Slack workspace conversations — when users ask the assistant to search or reference channel messages, that data is retrieved
in real-time from the Slack API and is discarded after the response is delivered. Customers can disconnect the integration at any
time, which triggers deletion of stored tokens and associated data.
Data archiving and removal policy
When a customer disconnects the Slack integration or deletes their Hubi account, all associated Slack OAuth tokens and AI
conversation history with the Hubi assistant are permanently deleted from our systems. Any remaining data is automatically purged
after 5 years from the date of account creation. Hubi does not store Slack workspace conversations (channel messages, DMs between
users, etc.) — only conversations with the Hubi assistant itself are retained. Customers can request data removal at any time by
contacting support or disconnecting the integration through their workspace settings. Data removal is completed within 30 days of
the request in accordance with GDPR requirements.
Data storage policy
Hubi, operated by Tidio, stores customer data in Supabase-hosted PostgreSQL databases located in the EU (AWS
eu-north-1 region). Slack OAuth tokens are stored encrypted at rest in the database. Conversations with the Hubi assistant are
stored in PostgreSQL for the duration of the subscription plus 5 years. Vector embeddings for memory features are stored in
PgVector (PostgreSQL extension) with the same security controls. Temporary session data is cached in Upstash Redis (encrypted in
transit and at rest). Hubi does not store Slack workspace conversations — channel messages and user DMs are fetched in real-time
from the Slack API during active sessions and are not persisted. All data storage practices comply with GDPR and applicable EU data
protection regulations.
Data center location(s)
Netherlands
App/service has sub-processors
yes
Guidelines for sub-processors
App/service uses large language models (LLM)
yes
LLM model(s) used
Claude Opus 4.6 (Anthropic), Gemini 2.5 Flash (Google)
LLM retention settings
Both Anthropic and Google APIs are configured with zero data retention. Inputs and outputs are not stored or used for model training. No Slack content or user data is retained by LLM providers after the response is delivered.
LLM data tenancy policy
Hubi uses Anthropic Claude and Google Gemini APIs. Each request is stateless and isolated per-customer. No data is shared across tenants. Only the context needed for the current conversation is sent to the LLM provider.
LLM data residency policy
Anthropic processes data in US-based infrastructure. Google Gemini processes data in Google Cloud infrastructure. All data is transmitted over TLS 1.2+ encrypted connections and is not stored by providers beyond the API request.