Your data
Adstract uses only the data required to run enhancement and reporting 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 analysis/reporting quality and request context integrity.
-
x_forwarded_for- Used as request-origin network context.
- Supports metadata generation and reporting context.
Data Adstract does not use
llm_response- The raw final model response content is not sent to backend reporting as-is.
- Reporting uses derived analytics and tracking metadata instead.
Why this matters
- Reduces unnecessary exposure of model output content.
- Supports privacy and data-protection expectations in production integrations.