Skip to main content

Your data

Adstract uses only the data required to run enhancement and acknowledgment flows.

Data Adstract uses​

  • prompt

    • Used as the source text for enhancement processing.
    • Required to generate the enhanced prompt output.

  • user_agent

    • Used to derive client metadata context.
    • Supports request context integrity.

  • user_ip

    • Used as request-origin network context.
    • Supports metadata generation and reporting context.

  • llm_response

    • The final model response text is sent to the backend as part of the acknowledgment flow so Adstract can analyze the response outcome.
    • Adstract does not save llm_response in any form.
    • It is used only for analysis during acknowledgment processing.

Optional targeting data​

When provided via OptionalContext, the following fields may also be used:

  • country, region, city — geographic targeting
  • asn — network context
  • age, gender — demographic targeting

These fields are entirely optional and are used only to improve ad relevance.

Why this matters​

  • Only the data required to run the integration flow is used.
  • Supports privacy and data-protection expectations in production integrations.