Skip to main content

Data flow

To understand how Qluent uses your data review the points below.

If you are accessing Qluent via app.qluent.com, the environment uses both GCP (Google Cloud Platform) and AWS (Amazon Web Services) to host and run the application. Custom client environments i.e. {client}.qluent.com are hosted on GCP resources only.

Qluent GCP only environment

The diagram above represents Qluent in a GCP only environment.

Caching​

Currently SQL queries produced by the LLM and data retrieved from your data source are not cached and reused in subsequent requests to ensure data freshness when you send a message/ask a question.

Storage​

Data stored by Qluent is encrypted at rest and uses a minimum of TLS 1.2 when in transit.

The following list illustrates data that is stored from each question you ask:

  • Conversation - the exchange of messages between user and Qluent
  • Intent - Qluents understanding of the question you are asking based on the conversation
  • SQL query - The query that was generated based on the intent
  • Data - The truncated dataset produced by the SQL query (can be configured in the web app)

To support each project the following data is stored:

  • Metrics/KPIs
  • Examples/validated queries
  • Context provided in the Knowledge
  • Columns that are enabled for semantic search

Cost control​

Qluent does not limit the user's ability to ask questions that yield large datasets. To maintain a good UX in the web app results are truncated to 1000 rows after retrieval from the data source.

If you would like to control query usage we recommend handling cost control at the client side with your database or data warehouse. For BigQuery this could be handled at user or project-level according to the docs